Home

PmSys: a monitoring system for sports athlete load, wellness & injury

1. Advocacy democracy Figure 2 8 Common architecture in participatory sensing applications 5 2 4 open mHealth Chronic diseases like diabetes asthma and obesity account for 46 of global diseases burden using traditional model for episodic care in clinic and hospital based settings is sub optimal for improving chronic disease outcomes 24 mHealth is based on the principles of eHealth which is health services and information with use of information and communication technology ICT mHealth move the focus towards the mobile phone instead 25 By using mobile devices in conjunction with Internet and social media we can enhance disease prevention and management by extending health interventions beyond traditional care this approach is referred in the article Open mHealth Architecture 24 The idea behind open mHealth is to make it feasible for patients to collect and share relevant data in real time to any time And not only when visiting a clinic it allow more rapid convergence to an optimal treatment For example a patient who needs to monitor their own health by writing notes for each day and report to the clinic monthly or weekly With an application to report this data in real time to a hospital or clinician who are following the patient s health status the patient can get instantly reports and the clinics can look for patterns of the response if any abnormal is present the patient can be contacted immediately for furt
2. Previous Submit Copyright 2015 UC Regents View Our Privacy Policy Survey is done Thank you v2 0 0 D CI al a Standalone version b MWF version Figure 3 11 Summary page in Ohmage applications 3 3 8 Conclusion Both applications did what we wanted them to do which was self reporting through a mobile application However with our evaluation we just mentioned the application could be improved and customized even more to fit an athlete better but also perform faster We could provide the user more useful feedback and allow the user to monitor and follow their own progression based on the reported data This makes the application not only a self reporting tool but also a monitoring application Hence we wanted to implement a new application that is customized and created to fit an athlete better we also want to have on application which look the same and provide the same features on all platforms We hope to replace both Ohmage applications with a pmSys mobile application this application will be presented in the next chapter Hence with the evaluations we just mentioned the application could be improved and customized to fit an athlete s need even better but also perform faster 3 4 Ohmage web portal The web portal in Ohmage is a web application that communicates with Ohmage backend server and fetches data from the server Ohmage portal is written in Java using Google s web framework GWT Google Web Toolkit and runs a
3. Figure 4 8 AngularUI state routes example 4 42 gt host 8081 visualizati v A localhost ses or HO gt C D localhost 8081 4 visualization Or MO Visualization aat Visualization Teams Demo Survey2 Players Wellness SIF Survey Players a b gt 1 localhost 8081 t visualization Or uo Visualization Last 30 days O Fatigue Sleep Soreness Stress Mood Sun 25 Jan c Figure 4 9 Different states same URL Main application Visualization Latest Last 30 days O Fatigue Sleep Soreness Stress O Mood Back to visualization Figure 4 10 Illustration of Angular UI parent and child view 4 3 2 Campaigns and surveys In Ohmage we have to manually download the campaigns we wanted to conduct and then it will be stored locally on the phone s storage Our application on the other hand handle this download 43 process differently by downloading all available campaigns automatically right after user have logged in Furthermore the campaings are stored locally on the file storage on the phone When the user want to report data on a later point the campaigns are always available The user can force application to download the latest campaign and survey updates from the server By downloading all campaigns and store 1f afterwards automatically does not need the user to manually download the campaign as we have to do on Ohmage The campaig
4. s mi se mt esnopseR en de 40 2 2 00 00 00 00 00 03 00 00 06 00 00 09 00 00 13 00 00 16 00 00 19 00 00 23 00 00 26 00 00 29 00 00 33 Elapsed time granularity 500 ms a Configuration read p E Amazon W Azure W Digital Ocean E UiT 6 000 5 400 4 800 4 200 3 600 3 000 i NN 2400 o c f 2 z j 1800 E v 1 200 600 e 0 00 00 00 00 01 00 00 02 00 00 03 00 00 04 00 f del 00 06 00 00 07 01 00 08 01 00 09 01 00 10 01 00 05 00 Elapsed time granularity 500 ms b Stateless authentication W Amazon W Azure W Digital Ocean E UiT 300 0 00 00 00 00 00 03 00 00 07 00 00 10 00 00 14 00 00 17 00 00 21 00 00 24 00 00 28 00 00 31 00 00 Elapsed time granularity 500 ms c Response read Figure 6 3 Latency on wired network 6 42 Mobile broadband network Latency through the mobile network was significantly lower on stateless authentication than wired network But configuration info had higher latency than wired network The most surprised findings was that Amazon had lower latency than Troms on stateless authentication In additional we could see that cloud services and University in Troms have slightly latency difference over the mobile network compared to wired network See Figure 6 4 and Table 6 3 for result TI M Amazon W Azure W Digital Ocean E UiT 2 000 1 800 1 600 1400 1200 1000 smi se 0 00 00
5. Object y campaign Object campaignName SIF Survey Players campaignUrn urn campaign sif players 2 eng v surveys Object y survey Array 3 gt 0 Object gt 1 Object gt 2 Object lenoth gt proto Array gt proto Object i proto__ Object E _ proto__ Object v Object campaign Object y campaign Object campaignName SIF Survey Coaches eng paign sif coaches 2 eng Parsed v Object campaign Object y campaign Object campaignName SIF Survey Players campaignUrn urn campaign sif players 2 eng y surveys Ubjec y survey Array 3 gt 0 Object gt 1 Object gt proto__ Array gt proto Object pb _ proto__ Object gt _proto__ Object v Object campaign Object y campaign Object campaignName SIF Survey Coaches eng campaignUrn urn campaign sif coaches 2 eng w surveys Object arpalgo a D a vw surveys Object survey Object y survey Array 1 1 0 Object b proto__ Object length gt proto Object Figure 4 11 X2JS converting XML to JSON 4 3 3 Report for yesterday RPE survey should be reported as closed as possible after ended session this is to get most accurate data on how an athlete actual felt about the session A problem we had with Ohmage application was when a player forgot to conduct their survey after a session but remembered to do so the day after With Ohmage they could not report the day after because the data will be s
6. Srivastava Participatory sensing In World Sensor Web 2006 23 Smartphone s in norway http www medienorge uib no statistikk medium ikt 379 Accessed 2015 04 18 24 Deborah Estrin and Ida Sim Open mhealth architecture An engine for health care innovation Sciencemag 330 759 760 November 2010 Link http research microsoft com en us um beijing events ms ubicompll mhealth science pdf Accessed 2015 03 12 25 mhealth for developers http www vitalwaveconsulting com pdf 2011 mHealth pdf 2011 Accessed 2015 04 18 26 mhealth getting started http www openmhealth org developers getting started Accessed 2015 04 18 27 Dhis user manual https www dhis2 org doc snapshot en user html dhis2 user manual en full html d5e30 Accessed 2015 03 01 28 Milan lab http www acmilan com en club milan lab Accessed 2014 10 12 29 Universitet 1 troms bruk av teknologi for prestasjonsoptimalisering 1 troms il http www olympiatoppen no om olt regioner regionnordnorge aktuelt media20277 media Accessed 2015 05 04 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 105 H Tangmunarunkit C K Hsieh J Jenkins C ketcham J Selsky F alquaddoomi D George J Kang Z Khalapyan B Longstaff S Nolen T Pham J Ooms N Ramanathan and D Estrin Ohmage A general extensible end toend participatory s
7. C 1 RPE Survey Rating of Session RPE When Within 15 minutes after each training session and match typically in the dressing room If that is not possible do it as soon as possible If you want to register the session you had yesterday please check for yesterday in the check box on the last page of the registration Registration contents Please answer the following questions 1 Was this a match team session or a individual session 2 Which type of session football session endurance session strength speed session or other 3 Duration number of minutes 4 How was your session today 0 10 What kind of information do we get Session RPE provide information on intensity duration and frequency of your training sessions and matches By tracking these data over time the coach can supervise the total training load and the variation in training load Clarifications e Dictionary Match official match or friendly match Team session Training session that include the team or part of the team Individual session Football session a session on the field which includes the ball Endurance session For example a running cycling or swimming session ment to improve endurance Strength training Strength training or core training typically in the gym Other Other activity that doesn t fit into the other e g playing tennis yoga e Rate how intense you experienced the session The rating should be an ave
8. Further I would like to thank Thuc Hoang and Cong Nguyen Nguyen for being great collaborators and this thesis would not been done without your creative ideas great discussions and hard working I personally want to thank both of you for the support motivation and encouragement throughout this thesis I would also like to thank H vard Wiig and Thor Einar Andersen from Norwegian school of sport sciences for giving me the opportunity to combine my thesis with your research in addition provided great feedback and discussions I also want to thank everyone that participated in the user study and everyone that have been testing and using this system throughout this thesis Next I would like to thank H vard Johansen Dag Johansen and Svein Arne Pettersen from University of Troms and Carsten Grizwodz and H kon Kvale Stensland from Simula Research Laboratory for your great discussions and inputs throughout this thesis Finally I would like to thank my friends and family for their support and encouraging words special thanks goes to my girlfriend Lina Mai Nguyen who have been a significant support to me throughout this thesis I could not finish this thesis without you Oslo May 18 2015 Kennet Vuong Chapter 1 Introduction 1 1 Background By monitoring a person over period we can learn about the person s unusual behaviors and habits A method for monitoring is to observe and capture relevant data for the person s activity T
9. Server Device s Distributal Browser Package Application Source Code Hybrid Framewrok SDK Tools SDK Tools Distributable Package App Store App Store User a Native application b Hybrid application c Web application Figure 4 2 Mobile application development approaches 6 4 2 2 Apache Cordova Apache Cordova is a hybrid mobile application framework that allows the developer to access native device function such as camera or accelrometer in Javascript Cordova applications are built with web technologies such as HTML CSS and Javascript Cordova provides application development without need to deal with any native code Java Objective c etc but only with use of web technologies Cordova is using the platforms SDK to wrap the developed application into a native container This makes the application able to be distributed through the platform s application store and use native features Cordova have large support for platforms such as 10S Android Blackberry Windows phone and more 35 All native features 1s accessed through use of plugins Cordova have already some basic plugins It exists more than 900 Cordova plugins 36 and the plugins are built in native code and it s platform dependent If we found a plugin we wanted to use we have to check whether the platform was supporting the platform we are developing for 4 2 3 Single Page Application Since Cordova is using web technologies
10. Server improvements 6 5 1 Authentication token One of the biggest struggle we had in the beginning with pmSys we wanted to change the expire timer for the authentication token The expiry timer for the authentication token was hardcoded into the server application see Figure 6 5 and sat to 15 minutes xx x User storage User objects are mapped to unique ids Avoids dependencies on x JEE session management The lifetime param set on construction controls how x long User objects stay active x Qauthor Joshua Selsky public final class UserBin xx x This is the length of an authentication token public static final int LIFETIME 1000 60 x 15 Source https github com ohmage server blob master src org ohmage cache UserBin java More Figure 6 5 Auth token timer was hardcoded to 15 minutes The reason we wanted to extend this timer is to relief the user and applications from refreshing the authentication token every 15 minutes When the timer expires for a token the user has to re authenticate to get a new token The only way to refresh the authentication token was to do a random request to the server with the authentication token and then the timer will be refreshed We moved the timer value from the source code into the preference table in the database When the server starts the authentication will be set by fetching the timer from the database With the new solution web applications and tokens is val
11. Structuring and organizing is yet another challenge in mobile application development especially when we work in team Model View Controller is an application structuring principle for implementing interactive application components 37 MVC consists of a three way layer to structure components 37 Model The model represents application domain state and behavior View The view deals with what the users see basically a graphical representation of the model Controller The controller handles the communication between the model and the view We have tried to follow this software application pattern in our application development UPDATES MANIPULATES VIEW CONTROLLER y Figure 4 4 MVC figure 8 4 2 5 User interface Due to our lack of graphical design experience and background using web technologies can ease some of the work needed to build a new application We can for example take advantage of the existence of various UI frameworks available to compose the application A well known UI framework is Bootstrap it provides pre designed components like buttons headers tables and more These components can be easily customized with use of CSS Bootstrap is created with mobile first principle which means the components in Bootstrap are optimized for mobile phones However the framework is designed to create web application running in the web browser rather than mobile phone only We want an UI framework that can offer native u
12. UnderscoreJS gt we extended the native JavaScript containers with new useful feature and functions for example sorting and finding various elements in big data collections MomentJS We used MomentJS to handle time and date in JavaScript this library allowed us to parse manipulate validate and displayed advanced formatted time stamps in JavaScript and the library Was very easy to use NVD3 is a library that is built upon D3 and offers advanced and reusable charts with web technologies such as HTML CSS and JavaScript We used this library to present all graphical charts ngTouch When developing a web application most of mobile web browsers wait about 300ms after tap and release before sending a click event 40 By using ng Touch library from AngularJS we can remove this delay and the click event will respond immediately Angular UI Router was another core library that was used in our new application AngularUI Router allowed us to translate URL routes into various states instead of writing long URLs 4 2 7 Conclusion Our new application was built with all presented theory design patterns technologies and frameworks together We decided to build a Hybrid mobile application due to our time limit and we needed to support Android and 1OS devices We used AngularJS as the core JavaScript framework in the application Ionic as the HTML and CSS framework to provide native UI We used Single Page Application and AJAX program
13. ae ee A 82 15 Nobile applicauon a a A A ide Ye a S TE DEG Fie 82 Paz MED PO sea ety et Oy Res ds de kr AA ek e Ge de 82 SIS see ek SEE Se se Fe Gage a G oH eS 83 Source code 85 Ad PSSS APD GL STIPLET STL LS TEEN ee ee ES 85 Aa POSTING as OR Se SRE DAGS TEN A SY Oe a A 85 Ad DOS PUSO sg La AS SSE Fe RSS GE EDGE Ge 85 Ad Improved Ohmag server 45 se sd As Ede Si e FR GN JG e ESA 85 5 ONE EN svake Pass RUG SE SEE FEE kle EE MG ee fre 85 User Surveys 87 B L pimisys Vs OMNdGe sae save FOT R DE AAA SE FR ORS 87 B2 Rate OF PMSYS Hdi ER RSA PTE LENE ER 91 Surveys in pmSys 95 or REMO ER A E A AS AAA A 95 C2 Wellness surveys 5 a a a GS O A id a ET 98 E TNS O E A SS A AS NE ERE ES 101 Vi List of Figures 1 1 21 22 23 2 4 25 2 6 2 7 2 8 2 9 2 10 2 11 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 3 10 3 11 3 12 3 13 3 14 3 15 3 16 4 1 4 2 4 3 4 4 4 5 4 6 4 7 4 8 4 9 4 10 4 11 4 12 4 13 GE SETET EE EEE A A RA 4 The EnvISONEA System paa G rd See GSE ds G gn 5 RPE Scale id 2 fo tb bate See ad e EE Ne ee eee 8 RPE related calc lati ns 2 c i osoro BR Ae hh ew SS La See Su 8 REE monitor ll y ses se ses a ee See A ee Eee a iS 9 An example of wellness sheet 3 o ran 9 SA E O EN 10 Injury questionnaire Av ds eee eee a ES BEE SEG 11 Common architecture in participatory sensing applications 5 12 open mH Se nace RED ETT
14. ale Goth Eee nn SAN Campa AA TAR ee ee ee F r E A DD ee Ge y ee LS Se S GSE SLT SET G Da ISCES SIM ar eee Ah AN EE be 3 3 Ohmage mobile applications rv rv vr ee JO Generalteaures 2 2 gere ee G3 he KA se Sud 32 Camparens and proeds ice ains tekkes Er KA Lals aa e sug FI GEER A AE AI IE 5504 Upload qUe gt 6 2 4 OSs A A A di A A ey A EE II 350 Response hon gt 2 2 md A A EGR 3 3 7 Evaluation of mobile application ee D CM 350 SN ot aie ds as ini a ios dd S Ohimace Web portal cz daa daria A a a Eg JAN CAMPAL OOS F R DAZ VEICWING TESPONSES sor sr ASS be F RE SEG 906 Ge dd 540 EXPO FSSpOnSES f 4 0 008 4 eh EO YS ORS DORR EER Ee KE GS S44 Nistalizauion sg 6424424464644 id 494 TG EEE SES 111 Se User ma a 10 1 y o id da e ds de do eo do 28 3 4 6 Hard coded endpoints o sesa acana aeee vr rn rn rare 29 SAT CONCIUSION s s ed Gee See SES AS a dd GP ket 29 GE 00000 EE rad eh EE VE NAN EN EE Ve SM Je 29 The pmSys mobile application 33 Al MOON Elo A e a A AA 33 T DE 0 a ad A n a O Gy E A a a BE a E A e 33 4 2 1 Mobile application development arr rv rn kr kr nr ra eee 33 ADD Apache Cordoba saa E A kr redere ss O me 36 223 Single Pace Applian 2 4 4 aaa 8 AAA LESE Le EG Gs 36 424 Software design pattem 62 o a ctw bee ele ee eee ee Se SS 37 425 Vermeer ass se PT ODS oe Pek eRe SPE es GaSe SPP oad 37 ADO Jaya TPE Lak ae cg cater Be eo de
15. and more They are offering mostly the same functionalities we want to provide where coaches can analyze the players reported data and the players can self report data through a mobile phone The problem with Athlete Monitoring system is that it cannot be easily extended by other developers and the system does not provide any open API like mHealth and DHIS2 The users of Athlete monitoring system is also bounded to use their software Another cons with Athlete Monitoring it comes with a cost and the more users we have the higher is the price What our envisioned system want to accomplish differently from Athlete monitoring 1s to have open public API where the system can easily be extended by new applications Then the users are not bound to use our software thus they can create their own analyzing tools and software 2 7 Ohmage Ohmage is a modular and extensible open source platform The platform provides mobile and web participatory sensing platforms that records stores analyzes and visualizes data from prompted self report and continuous streams 30 Ohmage is built upon the Participatory Sensing concept where the patient in health context can report different symptoms regularly and clinics can monitor their patients and provide instant feedback Ohmage have a rich set of API endpoints for their server back end that provides data access across all applications 30 Ohmage uses mobile application to collect data and a web portal to analyz
16. can only read their own responses but a coach can fetch and read all players data Other endpoints can be found in Ohmage s documentation 52 6 3 Cloud services Throughout this thesis we have been hosting our Ohmage server at University of Tromsg We never experienced any problems with their hosting through out thesis until the problems mentioned in Section 6 1 Under application development we did not think that the server could go offline Hence we wanted to explore and try cloud hosting and we think that problems we had with UIT or other self hosted environment can be solved with cloud hosting Cloud computing is hardware systems software and application delivered as a service over Internet 53 In other words cloud computing is just another computer service manged over the internet One of the benefits of picking cloud providers to host our ohmage server application 1s that most of the cloud services can provide 99 uptime Another beneficial with using cloud services they offers scalability which means we can scale our system and increase hardware specifications to our needs Let the cloud service handle infrastructure and hardware upgrade might even be more cost efficient than self hosting Within cloud computing we have three cloud service models 54 e Software as a Service SaaS where we can deploy our application to the provider s cloud infrastructure but we can not control anything else like operating system hard
17. charts In addition to store and process reported data have a persistent distributed server 1 3 Limitations As we mentioned in our problem statement and in this thesis we will focus on collecting subjective data through questionnaires and self reporting We want the system to support third part objective measurement devices in the future but in this thesis we have limited our scope to only collect subjective data Our goal are to develop a modern and digitized tools to collect process store present and analyze therefore we prioritize usability and functioning tools higher than user interface design Other technology possibilities such as integrate third party devices machine learning and artificial intelligent 1s not a part of this thesis but should be considered as future work Since we are trying to support NIH s research with our envisioned system NIH will be our test users and provide feedback on our research Hence the system has only been tested in football context and not in other sports contexts Security aspects and challenges of the system will not be taken to consideration in this thesis but we have partly taken this to account by storing encrypted user ID s Thus there is no way to identify the person We wanted to test the system in various cloud services with the primary goal to find any latency advantage from one cloud service compared to another We will not perform any in depth research to the network subject More
18. focus over to the server The server 1s the place where all data 1s stored and processed One of the reasons we wanted to use Ohmage and built our system around it was because Ohmage provided a server backend that could handle reported surveys but also present surveys and captured data We could just build applications to work around Ohmage Ohmage also offers a great user management system with clear user roles that help us separate a player role from a coach role and keep data from unauthorized users When the pmSys applications were built we saw that Ohmage server provided unnecessary data and the application needed to do many unnecessary server requests to get enough data before we could visualize and present it We started to look at the server implementation and wanted to find optimizations and improvements towards the server Throughout the thesis we hosted our Ohmage server at University of Troms UIT They were responsible for maintaining and keep the server secured We never experienced any problems with their hosting Until it came to a point where the servers at Troms needed an upgrade UiT performed a hardware and UPS uninterruptible power supply upgrade and the server was taken down for few hours We had another scenario where the electricity at Tromsg went down and the server were unavailable for the days We did not consider that the server could go offline and neither did we have any backup plan Due to server offline pla
19. in Figure 3 1 Ohmage front end Q ohmage Web portal Ohmage server Q ohmage Backend server Data collection applications Ohmage Standalone Ohmage MWF Android iOS Figure 3 1 First version of pmSys using Ohmage 3 2 Ohmage system We got a brief introduction to Ohmage in Section 2 7 Ohmage platform consists of four primary components 1 Ohmage backend that serves as data storage and provide unified interface for data access 2 Mobile data collection applications that runs on participants phones for collection and feedback 3 Web based data management and administration tools for study management and administration 4 Web based data analysis and visualization tools for exploring analyzing and visualizing the collected data 30 Ohmage fulfilled our requirements as a system and provided a system that could report process and present data Since the system was open and extendable with new applications we could develop new application to fulfill our usage if necessary The first version of pmSys was based on Ohmage system Moreover we used all Ohmage applications such as Ohmage mobile application for self reporting and the front end web portal to analyze and present data We distributed Ohmage mobile application to our users for them to start report Next the coaches were introduced to Ohmage web portal where they could view captured data In the next sections we present some features Ohmage system was providi
20. irritable down Figure 2 5 An example of wellness sheet 3 10 2 2 3 Monitoring illness and injury In the paper by Oslo Sports Trauma Research Centre a new approach to prospective monitoring of illness and injury in elite athletes 19 they introduced a method for monitoring players based on weekly questionnaires The surveys were sent to athletes over email weekly After an athlete has finished the survey a project coordinator collects them and assembles a report based on the responses Then coordinator sends the report to medical staff which follows up with the athlete 19 An example of this monitoring process can be found in Figure 2 6 a However 1f an athlete forgot to report the medical staff had to send a reminder over email The weekly questions are to capture health problems such as injury and illness for a player The survey starts with four key questions which can be found in Figure 2 7 If the athlete choose the minimal value for each questions no problems the questionnaire is finished for the week 19 However if the athlete reported anything else than the minimum value the questionnaire continues by asking them whether their problem is an illness or an injury If the athlete reported an injury problem the following question is where the injury is located and if it was an illness problem the athlete was asked for symptoms see Figure 2 6 b Based on the four key questions each answers form each question is
21. m3 medium 1 3 3 75 1x4 SSD 0 077 per Hour m3 large 2 6 5 7 5 1 x 32 SSD 0 154 per Hour m3 xlarge 4 13 15 2 x 40 SSD 0 308 per Hour m3 2xlarge 8 26 30 2 x 80 SSD 0 616 per Hour a Amazon EC2 Basic tier prices 62 BizTalk Sener Oracle Sofware REGION CURRENCY Central US v Euro General purpose compute Basic tier An economical option for development workloads test servers and other applications that don t require load balancing auto scaling or memory intensive virtual machines INSTANCE CORES RAM DISK SIZES PRICE EN 1 0 75 GB 20GB 00135 hr 10 mo EN 0 0351 hr 27 mo EVI 2 3 5 GB 60 GB 0 0701 hr 53 mo 4 7 GB 120 GB 0 1401 hr 105 mo A4 8 14 GB 240 GB 02801 hr 209 mo b Microsoft Basic tier for Linux hosting prices 63 Simple Pricing All plans are standard with solid state drives SSD MONTHLY Oy HOURLY Additional bandwidth transfer is only 24 per GB 5 mo 10 imo 20 mo 40 mo 80 mo Most Popular Plan 512MB men 2GB Memory 4GB memory 8GB Memory 1 Core Processor 1GB Memory 2 Core Processor 2 Core Processor 4 Core processo 20GB SSD pisk 1 Core Processor 40GB SSD Disk 60GB SSD Disk 80GB SSD Disk 1TB Transfer 30GB ssp Disk 3TB Transfe 4TB Transfe 5TB Transfer 2TB Transfer SIGN UP SIGN UP SIGN UP SIGN UP SIGN UP c Digital Ocean prices 64 Figure 6 2 Cloud services 76 M Amazon W Azure W Digital Ocean Ei UiT 120 A l
22. minimum number is 5 and maximum is 300 When we click on plus sign the number is incremented by 5 to get from 90 to 5 only requires 17 clicks 85 5 In ohmage this process requires 85 clicks The timestamp prompt in pmSys is nothing different from Ohmage s timestamp sRPE after session sRPE after session Planned sRPE Question 4 4 Question 3 4 Question 1 4 How was your session today Enter duration in minutes Date and start time for the session O Rest gt EN 90 Date Mar 31 2015 1 Very very easy gt Time 4 59 PM 2 Easy gt 3 Moderate gt 4 Somewhat hard gt 5 Hard gt 6 gt 7 Very hard gt 8 gt gt Clear Done 9 gt 10 Maximal gt January 29 2013 February 30 2014 March 31 2015 April 1 2016 May 2 2017 a Single choice prompt b Number prompt c Timestamp prompt Figure 4 14 Supported prompts in pmSys mobile application 4 3 6 Extra presentations For each question presented in the application we have created a number indicator to show how many questions we have left of the survey and how many we have answered This indicator is presented differently in Ohmage standalone and did not exist in MWF version on 10S PmSys s presentation gives the player a better indication on how many questions they have left the result is presented in Figure 4 15 After we have finished a survey both Ohmage applications shows a summary text which is defined in the campaign file on u
23. o i SE EN EE EE EE a a 60 SL JENSON arv Se EVE RE AS Ne See 60 Sn REP ve NA SE Se LIA Se Su 61 OA ViSUdlIZaUON 44 da rd A A A Ge 62 3 3 3 Push Messages for coaches s s s bi ec ek AA A ti ee 64 5 3 6 Pmsys Push message service ais eRe a FASTE A 64 A Evaluation ez dISCUSSION Labs pd E a a GS rd amp EEG dd Sek 66 5 4 1 Visualization bottleneck a 66 542 PUSH message SENG Lade dd Ad BES Gb 66 5 4 3 pmSys vs Ohmage frontend 0 0 00 a 67 5 3 SUMMA bo sk ee kile A A de SE kN SE eee eS 68 The pmSys backend server 71 CE io AENA AN 71 02 OI SNE Sc e pae a ree Be ee de eee hee ere 71 6 3 Cloud services s os co EE AES ts EE dd de ee 73 GJ CId Provides e seia n kan bods oS LEG KSS i oe Re we d d 73 6 4 Cloud hosting latency vr rv a a G p a era rn 22 74 OAL WEG NCIWOPK 4 sa ra e is a la Se SS es 74 6 4 2 Mobile broadband network e e eo 76 64 3 CONCIUSION 2 04524 SPE Bee a eee A E NE el 65 Sererimprovemens ss 2 4 26 4 e oss T FAN SATS da da EG ge 78 6 5 1 AUthenticatiOMtOKEN a e tr ea Se AOE Eee AA ee eRe ee 78 6 0 2 NAMSEN aa ee eee Pa ee EE AA RES 78 655 Improved data fetching gt lt roe reni wa Oe eae eee SES SS 79 634A Conclusi n gas fars sort ORS of PRR SRY SSE SES EHO CRED OSS 79 06 o AO IA hs S harass PO Eg 79 Conclusion 81 PAL SAN E A A A SERT TG he de da 81 72 Man Conrbutons sus akv afarer A a O AER Ga a SG 81 des FUU Wok 21d ON EE EE as A A A
24. prompts is the same whether the internet is present or not However the only thing we have done differently from Ohmage 1s to upload reported surveys that are stored offline automatically when the application have or come in range for internet connection On Ohmage the user must upload the responses manually 4 3 9 Reminder Reminder feature was an great feature to help players report in Ohmage In pmSys we have only support for time based reminders and not location based reminders like Ohmage In Ohmage applications we could assign a survey to a reminder but in pmSys we have decided to left this out and thus we provide more simple reminders The user can set a time for when they want a reminder to be triggered and the message to be shown but also whether the reminder should repeat it self For this have we used a Cordova plugin local notification which allows us to schedule a notification nativly on the platform Mnttps github com katzer cordova plugin local notifications 47 A technical challenge we had with this plugin was the API functions this plugin was providing was not consistent The API did not provide any option to get a list of scheduled notifications and the information of triggered notifications where not providing correct information So we have to create our own containers for the reminders We have to store reminder information at three different places and keep these three containers synchronized We first save
25. single connection thread The test were done on an iPhone 4S with Telenor Mobile Network using 3G We measured the latency for three different Ohmage endpoints 1 Server configuration information 2 Stateless authentication 3 Fetching one RPE response from a test user 6 4 1 Wired network From wired network we see that Tromsg provided outstanding and lowest latency The primary goal with this study was to compare the cloud services and have Tromsg as a basis point Amazon and Digital Ocean provided lower average latency than Azure overall Digital Ocean have lower latency on configuration read But on stateless authentication which is more important for our applications Amazon provided lower latency than the others For our system this means the users can authenticate faster For reading responses were both Amazon and Digital ocean very close on providing lowest latency See Figure 6 3 and Table 6 2 for result Configuration read Stateless authentication 2087 ms 1503 ms Digital Ocean 3680 ms University of Troms Table 6 2 Average latency on wired network Fornebu Norway 75 Linux RHEL SLES Windows Windows with SQL Standard Windows with SQL Web Region EU Ireland 9 vCPU ECU Memory GiB Instance Storage GB Linux UNIX Usage General Purpose Current Generation t2 micro 1 Variable 1 EBS Only 0 014 per Hour t2 small 1 Variable 2 EBS Only 0 028 per Hour t2 medium 2 Variable 4 EBS Only 0 056 per Hour
26. they would not be able to read or analyze any data The purpose with this coordinator role is that they can analyze and interpret data for multiple teams An example for this role 1s the researches and coordinators from NIH We are using Ohmage classes to define a team within a team we have two different classes one class that represent players and the other class represents coaches Each class in Ohmage has an uniform resource name URN This urn is used as a unique identifier for a team and the syntax is urn class nameOfClass In pmSys we have included the team name s initials in the urn and the group role which are coaches or players At the end of the class URN we added a number this 1s to separate multiple teams within a club or a club with same initials An example of team demo s URN would be urn class demo players 1 and for the coaches 1t would be urn class demo coaches 1 If the demo club has another team user group and want the data isolated from team 1 we can add a class with the URN urn class demo players 2 and still have the coaches accessing data for both teams An illustration of pmSys user system with Ohmage is presented in Figure 3 3 This figures illustrates two example teams alfa and beta The two teams have two classes which represents the coaches and the players The coaches are privileged in their class and the players have restricted privilege The player class 1s assigned to the player campaign where the su
27. to build an application we use the single page application approach to create our application According to Single Page Web Application by Mikowski and Powell 7 An Single Page Application SPA is an application delivered to the browser that does not reload the page during use by reload means that the page is not refreshed on user interaction This approach gives the user a native feeling of using the application Also avoids view rendering on server side and we move the view rendering over to client side By doing this we are reducing the data transferred between the client and the server and let the server focus on sending relevant data rather than how buttons and other component should look like Figure 4 3 illustrates how SPA application works in theory and how the business logic and HTML rendering is transferred to client side and the server is focusing on sending and receive data Database Server Client Authentication Data storage Authorization HTML rendering Data retrieval Validation i Decorative JS Business logic Business logic HTML generation Traditional Database Server Client HTML rendering Data st Authentication Decorative JS Data retrieval oo OE Business logic JS HTML generation SPA Validation Figure 4 3 Tradional web application vs Single Page application from Manning Page 8 7 37 4 2 4 Software design pattern Building mobile application is challenging in general
28. we going to get a brief touch on DHIS2 which 1s a health information system developed by University of Oslo Next we talk about how other football clubs monitor their players and another existing monitoring system At last we present the most relevant system for pmSys We summarize and conclude with a brief evaluation of presented systems 2 1 System requirements In this section we present general requirements for our envision pmSys The system should provide a way to capture and submit data through self reporting Then the system must be able to process and analyze submitted data At last the data captured data must be persistent The system should present the collected data through visualization with user friendly and interactive graphical charts this allows the users to monitor and evaluate captured data The envision system is illustrated in Figure 2 1 We distinguish between function requirements and non functional requirements The function requirements describe the functions the system should be able to do Non functional requirements describe how the system should work behave and perform Distributed server Fetching data Sending data User Player User Coach Figure 2 1 The envisoned system 2 1 1 Functional requirements Self reporting We wanted to use the mobile phone to collect self reported data with questionnaires Therefore the mobile phone must provide a way to present and submit finished responses for example
29. weekly load chart We present three different charts plotted onto one graph load monotony and strain Visualization Visualization Visualization RPE Load Weekly RPE Load Weekly RPE Load Weekly 10 536 OG Load Strain Monotony right axis 1169 0 7 500 1100 0 70 9 450 1000 0 65 8 0 60 400 900 0 55 7 350 209 0 50 el 700 0 45 300 0 40 5 600 250 0 35 500 4 0 30 200 400 0 25 3 150 300 0 20 2 400 0 15 200 0 10 i 50 100 0 05 0 0 0 0 0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 15 16 17 18 19 20 21 22 23 24 25 26 27 28 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Back to visualization Back to visualization Back to visualization a Weekly rpe score b Weekly load c Weekly summary Figure 4 19 RPE visualization in pmSys 2Montony Weekly load Standard Deviation of weekly load 3 Strain load monotony 49 For wellness visualization we have created two different charts see Figure 4 20 The first one is a Histogram chart showing the latest reported score for our wellness data On the second chart is a Line chart providing wellness scores thirty days back in time in additional we have calculated the average score for the wellness data The last visualization feature we have implemented is a player s injury severity score calculated and presented with three months back in time See Figure 4 20 c Visualization Visualization Visualization O Fatigue Sleep Soreness Stress O Fatig
30. with new endpoints to reduce latency and data transfer between the client and the server 69 Campaign SIF Survey Players Survey All Ke Players Participant All i SLIWarzu Privacy sif 2cvdop sif 1396v0 Start Date sif jxrokk sif zhkf16 e sif liso7g End Date sif G2ipsi sif 0w616j y sif uzti6m Only photo responses e sif ai3la9 n sif 7ylan0 Show Responses e sif r chyg sif n9czhb sif qixot sif tdsw99 sif ryqotz sif 32hody sif un6qw5 sif 4nriwd sif 2bz8jw sif ueukjh sif bq7l0e sif evrixa a Ohmage response filter b Player list Figure 5 18 Responses in Ohmage and pmSys pa PMSys e pmSys Web portal pmSys portal pmSys Mobile application ohmage Ohmage server iOS and Android Figure 5 19 pmSys with new portal 70 Chapter 6 The pmSys backend server In this chapter we present technologies used in Ohmage server And then we study and analyze whether we can find any advantages of hosting Ohmage server in cloud with focus on Latency We only analyze network latency to the various cloud services and compare it to our self hosted environment in Troms At last we present possible improvements to Ohmage server and how these improvements can reduce the latency and data transfer between the server and client 6 1 Motivation Until now we have presented our implemented pmSys mobile application and web portal In this chapter we move the
31. 00 00 00 00 00 00 01 00 00 02 00 00 03 00 00 05 00 00 06 00 00 07 00 00 08 00 00 09 00 00 04 Elapsed time granularity 500 ms a Configuration read Mi Amazon W Azure W Digital Ocean E UiT 2250 2 000 1750 1500 0 00 00 00 00 00 02 00 00 05 00 00 08 00 00 10 00 00 16 00 00 18 00 00 21 00 00 24 00 00 27 00 00 13 Elapsed time granularity 500 ms b Stateless authentication MAmazon MAzure M Digital Ocean E UiT 00 00 00 00 00 00 00 00 01 00 00 01 00 00 02 00 00 03 00 00 03 00 00 04 00 00 04 00 00 05 00 00 02 Elapsed time granularity 500 ms c Response read Figure 6 4 Latency on mobile broadband network 6 4 3 Conclusion Our goal with the latency measure was to see which of these services could provide lowest latency from a wired network access but also from a limited mobile broadband network Since most of our users are using their mobile phone to report data providing them low latency make the system in general faster Based on the results we conclude that cloud service is a good solution when we are reliant latency In additional cloud services can guarantee up to 99 up time University of Tromsg provided the best latency from wired network but 1f we had to pick a cloud service would Amazon the one who provided 78 Configuration read Stateless authentication Digital Ocean University of Troms Table 6 3 Average latency from Mobile network lowest latency overall 6 5
32. 07g Msif lwa7z0 Msif ai3la9 MW sif r chyg Msif jxrokk Mi sif eluewO Msif Ow616j W sif 7ylan0 Msif n9czhb Msif h2dfog Msif tdsw99 Mi sif 2bz8jw Wi sif zxy00u Mi sif zhkf16 Mi sif 1396v0 Msif uzt6m MM sif bq7l0e Wi sif roscip Mi sif qixqt Mi sif 2cvdop Mi sif 62ipsi Msif rv23ie Msif 32hody W sif un6qw5 Msif 4nriwd W sif t7qif3 MM sif a0293z a Average wellness visualization Figure 5 10 Example of wellness visualization lul Injury 110 100 85838388 o gt n i Z gt m p ing E i lt lt i A gt lt _ E a gt an e be i Sa T T T T T T T T T T T T 27 03 2015 29 03 2015 10 04 201804 2015 18 04 2015 24 04 205M4 2015 03 05 2015 08 05 20 203 f FA T T T T T T T T T T T T T T T T ol 27 03 2015 29 03 2015 10 04 201804 2015 18 04 2015 24 04 283D4 2015 03 05 2015 08 05 20 Wsif iiso7g Msif lwa7z0 W sif aidla9 MW sif r chyg Mi sif jxrokk Mi sif eluewO Mi sif 0w616j ME sif 7ylan0 Ml sif n9czhb Msif h2dbg Msif tdsw99 Mi sif 2bz8jw Mi sif zxyOou Mi sif zhkf16 Mi sif 1396v0 Mi sif uzti 6m E sif oq7l0e Msif roscip W sif qixqt E sif 2cvdop Mi sif 62ipsi Mi sif v231e Mi sif 32hody Mi sif un6qw5 Msif 4nriwd W sif a0zq3z a Severity score Figure 5 11 Team visualization of Injury Individual visualization The coach can also visualize and monitor each player individually Here is all visualization features presented on one page A
33. 2 04 2015 17 04 2015 22 04 2015 E sit un cqws a Highlight single player lil SRPE Team sessions Coach expected 2 Ml Team average 2 E sif 1s07g 2 BB sif ai3la9 2 E DIN N H sif jxrokk 20 i W sif eluew0 2 E sif Ow616j 2 A sif 7ylan0 3 ee ME sif n2dfbg 23 03 2015 E sif tdsw99 1 1015 08 04 2015 12 04 2015 17 04 2015 22 04 2015 Coach expected Mi Tean l sif 2bz8jw 2 sif iwa7zO Msif ai3lag Mi sif r chyg Msif jxrokk El sif eluewO Mi sif 0w616j Ni sif 7ylano Misif n9czhb Msif h2dfog Mi sif t E si 2 Msif roscip Mi sif qixqt Misif 2cvdop Msif rv23te Msit 32hody W sif un6qw5 Msif 4nriwd Mi sif t7qi GJ RE gt ZINE SRS Ay B zm va re Aa er y IN JAN NA ANS N la FA 1015 08 04 2015 12 04 2015 17 04 2015 22 04 2015 b Data on date Figure 5 8 Interactive graphs Ll SRPE Team sessions RPE Load mod 600 1 Y 4004 ES YN h eel KN IS HA so PNL AZ A at ar EEE 04 i r T T T E T T EN SE T T 28 03 2015 31 03 2D1 84 2015 08 04 2015 12 04 2015 17 04 2015 22 04 2015 27 04 2015 02 05 2015 07 05 2015 E Coach expected Wi Team average Wi sif liso7g Msif iwa7z0 Msif ai3lag Mi sifr chyg Wi sif jxrokk MM sif eluewO Mi sif Ow616j Wi sif 7ylano Mi sif n9czhb Wi sif h2dfog Msiftdsw99 Wi sif 2bz8jw Misif zxyOou W sif zhkf16 Msif 1396v0 Mi sif uzt6m NM sif bq7l0e Misif rosclp Misif qixat Misif 2
34. 70 ED 1 1 33 WED APRIL 1 5 Message from kkvcoach 7 omSys o Remember to do sRPE Today DEE update available ndroid 5 0 1 System Update pmSys You have forgot to do sRPE Mm USB debugging connected ch to disable USB debugging Y Connected as a media device ouch for other USB options 7 pmSys You have forgot to do sRPE b iOS c Android Figure 4 18 Push messages sent from a coach 4 3 11 Visualization Ohmage standalone had a great feature where the user was able to view their response history on data they have submitted to the server We wanted to provide our users some feedback but better and more useful feedback This will also motivate the players do self reporting more often In pmSys mobile application we visualize reported data from the players and present it on various charts This allows the players to monitor their weekly load wellness and injury history All visualization charts are based on theory from the articles we presented and discussed in Section 2 2 Our visualization feature is the best and most relevant feature for the athletes since they can monitor their own progression In Figure 4 19 we can see all three RPE charts presented in pmSys The RPE visualization only shows two weeks back in time tThe first one is plain RPE score presented in a graphical chart The second chart shows the load score which is the reported session duration multiplied by reported RPE score The last one is
35. C MEUS ee akse EE ETE SS Sa S r 37 Ajax programming model 9 e 38 Module based development 22 aa vr vr arr knarr ren 40 Example of routing in AngularJS e 41 AngularUl state routes example e 41 Different states same URL sa sd vase a Sh eR a we AAC Mee Rake SEG 42 Illustration of Angular UI parent and child view 00 42 X21S convertie XML to ISO N aacra s aa lc a rs ds la ee dt eS 43 Summary page with option to report for yesterday 22 rv rv rv vr 22 ra 44 Reporting injury lt ves base GR A SPT TARE SE SEE EEG 44 Vil Vill 4 14 4 15 4 16 4 17 4 18 4 19 4 20 4 21 4 22 4 23 4 24 4 25 4 26 4 27 5 1 52 di 5 4 5 5 5 6 Jel 5 8 59 5 10 5 11 5 12 5 13 5 14 5 15 5 16 5 17 5 18 5 19 6 1 6 2 6 3 6 4 6 5 6 6 6 7 6 8 Supported prompts in pmSys mobile application 2 2 20000 45 Presenting how many question s left rn 45 Example of pmSys file system rn rn kr nr rar an 46 Reminder in PMSYS e c ca 244 sk a dees de AEE GS ES SEG 47 Push messages sent from a coach aar vr evna kr aar ren 48 RPE visualization inpmSys 222 rv va vr vrak rer rare 48 Nisualization OL Wellness server re a Siete s ee SSE karer 49 GPS LPIS Sos EAS AER SARA SEE ke S EA KE se SJ 50 Result of user study Ohmage vs pmSys vr rv rn nrk rn nr ren 32 Application p
36. E SKEE Eee de 13 DETS PIATOM ae e hor skr es Skeet L k SS Ge ee Oy Se dass eg 14 Ohmage architecture with applications built around Ohmage backend 15 First version of pmSys using Ohmage 2 00 2 eee eee 18 A tiny example of a Campaign file 2 ra ee 19 PIS YS USF SNE g o A e a ee A e e la e de G 21 Not able to select server on Ohmage MWF Android version 22 Screenshot of Ohmage dashboard screens vr rer e 23 Adding a campaign on both applications 2 vr nrk kr re rare 23 Surveys and prompts inpmSys LL 22 vr kr rv rn rare 24 Upload queue in Ohmage 22 av rv rav varar rn narr a a rare 25 Adding f mind rs 10 Ohmage ss 2545 b e d wehbe DOG Me baw ohh did baka 25 Response history in Ohmage standalone rer ra 26 Summary page in Ohmage applications rn rn rn e 27 An example of viewing responses through the web portal 28 Leaderboard in web portal e 29 Visualization in web portal 44 s id wh ee GE EEE a we oe SG abs 30 Class administration in web portal nr ran 30 Hardcoded Front end endpoints vr ere ar nr rare 31 Mobile operating systems release 6 LL arr vr rn nrk ran 34 Mobile application development approaches 6 222 aar arr arr ra 36 Tradional web application vs Single Page application from Manning Page 8 7 36 MV
37. I let us work on states instead long URL s We only need to specify various states and resources and the library will handle requested resources An example from our application where we used AngularUI routes to define various routes a user can access 1s presented in Figure 4 8 In Figure 4 9 we see an example of three different state and how the URL is still unchanged Another example of AngularUI is how the three states menu teamSearch and wellness is inheriting from the parent view visualization in Figure 4 10 The state module is working like a tree structure Example taken from https scotch io tutorials single page apps with angularjs routing and templati Accessed 2015 03 30 create the module and name it scotchApp also include ngRoute for all our routing needs var scotchApp angular module scotchApp ngRoute ff configure our routes scotchApp config function routeProvider SrouteProvider route for the home page when 4 templateUrl pages home html controller mainController route for the about page when about templateUrl pages about html controller aboutLontroller route for the contact page when contact I templateUrl pages contact html controller contactController HD IG create the controller and inject Angular s scope scotchApp controller mainController function scope create a message to display i
38. RPE Team sessions pmSys portal 2 ohmage Ohmage server Figure 5 6 An overview of pmSys portal 5 3 3 Responses How the web portal presents the responses is differently from Ohmage web portal instead of a filter to view the responses like Ohmage provides We first present a list of accessible campaigns secondly we present a list of available surveys before pick a survey we present a statics over reported data the last 14 days See Figure 5 7 a And after the user have picked a date to view responses the portal presents a list over who has reported that day and who have not See Figure 5 7 b And then the response 1s presented to the user in Figure 5 7 c 62 EiSurveys Surveys Survey name Description sRPE after session Rating of training load Wellness every morning Rating of wellness Injuries lliness once a week Injury illness registration Statistics of responses last 14 days 3 25 E 15 Apr 17 Apr 19 Apr 21 Apr 23 Apr 25 Apr 27 Apr 29 Apr E sRPE W Wellness W Injury a Available surveys statistics Players Responded sif qixat Sif r chyg sif 0w616 i i Ej Player response sif zxyOou sif h2dfbg Player sif eluewO Reponsetime 2015 04 29 20 17 54 sifalla9 i ae Question Answer Sif Jxrokk Not re spon ded Choose an event Match sif iwarzO Choose type of session NOT DISPLAYED sif rygofz Enter duration in minutes 35 sif 32hody sif un6qwd How wa
39. U10 Department of Informatics University of Oslo PmSys a monitoring system for sports athlete load wellness amp injury monitoring Player monitoring system Kennet Vuong Master s Thesis Spring 2015 PmSys a monitoring system for sports athlete load wellness amp injury monitoring Kennet Vuong Abstract Most football players mostly young talents end their bright future and football career due to injuries and overloads A player s performance during a game might be affected by various factors such as muscle soreness fatigue mood or sleep quality By monitoring a player we may be able to predict and detect a possible injury or prevent a reduced performance in a game The most common monitoring tools used in sports context today are still pen and paper In this thesis we present an implemented digitized monitoring system where we are monitoring a player s load wellness and injury through questionnaires The system has made it easier to collect process store analyze and present data using modern technologies We have used a mobile application to collect data the application also provides useful visualization feedback thus helps a player to monitor them self The system also has a web portal where a coach can analyze evaluate and monitor players captured data This player monitoring system pmSys can help sports clubs and athletes to ease the process of manually capturing and processing data and rather have foc
40. ach answers are located Based on the performance test results in Table 4 1 and Figure 4 25 is our application faster when it comes to conducting a survey pmSys Avg Seconds Ohmage Avg Seconds pmSys Clicks Ohmage Clicks PRPE 7 052 12 498 n EEC 5 945 ug e gt 4 384 6 103 e mw 2313 15 543 Table 4 1 Performance test pmSys vs Ohmage 4 4 5 Conclusion The whole purpose of building a new application was to replace Ohmage s mobile applications and see if we could create a improved and customized application to better fit an athlete We have optimized the application with new features and provide more useful feedback than Ohmage We have removed unnecessary user interactions and features and created a faster and useful reporting tool for the sports context and athletes We have developed an application that is distributed on two different platforms but provide the same functionalities on both platforms Being able to receive push messages give the user possibility to have a unidirectional communication with the coach With the new application have increased the number of responses tremendously With the four clubs from the Norwegian top division which used Ohmage application in the first version of pmSys was total collected responses only 467 and this was numbers for the whole 2014 season from four different teams With the new application pmSys the total responses from two top division teams are over 5000 responses in t
41. acy Policy v2 0 0 v2 0 0 lt D cA al a Empty project list Android b Empty project list GOS c List of available campaigns 1OS Figure 3 6 Adding a campaign on both applications 3 3 3 Surveys When we click on the survey function we get to choose between surveys we want to conduct user can choose between RPE Wellness or injury report Both Ohmage applications support most of the prompt types but within the three surveys we have presented we have only used single choice number and timestamp The method for reporting a survey were the same on both applications we pick a survey and are presented a question and various options After we have picked the option we wanted we click the next button to move to next question When the survey is finished we get presented a summary text On Ohmage MWF we were asked whether we want to share our GPS location or not and the users were told to click no here Because we do not want to save any data that could be used to identify a person in real life Screenshots of the prompt types and available surveys are presented in Figure 3 7 24 6 i hot I eeN Telenor 18 o KE Q sRPE after session Q sRPE after session Available Surveys sRPE after session Match Rating of training load Wellness every morning Team session gt Rating of wellness Injuries lliness once a week Injury illness registration Upload Queue Copyright O 2015 UC Regents View Our Priva
42. alth problems To what extent have you reduced you training volume due to injury illness or other health problems during the past week To a minor extent To what extent has injury illness or other health problems affected your performance during the past week To a major extent To what extent have you experienced symptoms health complaints during the past week To a severe extent Is the health problem referred to in the four questions above an injury or an illness Illness What is the illness related to Stomach region Please state the number of days over the past 7 day period that you have had to completely miss training or competition due to this problem 1 Is this the first time you have registered this problem Yes first time Have you reported the injury illness to the medical team No ave you experienced any other illnesses injuries or other health problems during the past 7 days Yes Take part 2 Back to program Injuries lliness once a week Result success Have you had any difficulties participating in normal training and competition due to injury illness or other health problems during the past week Full participation without health problems To what extent have you reduced you training volume due to injury illness or other health problems during the past week To a minor extent To what extent has injury illness or other health problems affected your performance during the past week To a minor extent To what extent
43. and distribute the application to the store Hybrid application saves a lot of time in the early stages because we have to deal with only one language but as the application escalate and become bigger we have to deal with the application separately by device in the end 4 42 User study Ohmage Mobile Application vs pmSys Mobile Application We performed an user study to confirm whether our application was more user friendly and better than Ohmage We invited 26 persons in age s from 16 to 33 to perform the study We started with presenting the purpose of the study without telling which application we have created and then we told the users to put their self in an athlete s perspective before we started to test the application We started with Ohmage mobile applications on both Android and Phone first Then we gave everyone an username and a password and told them to log on and try to perform a survey in the application and try to navigate around the application and use it in general After users were done with testing of Ohmage applications we moved to our application pmSys Here they did the same procedure of logging in conducting a survey and navigate through the application We did not tell them which application was ours and neither did we give them any hints or directed their thoughts to prefer pmSys over Ohmage After testing both application we provided them a survey where we asked for their feedback on S usability us
44. articipate at all Question 3 To what extent has injury illness or other health problems affected your performance during the past week O No effect O To a minor extent Q To a moderate extent O To a major extent O Cannot participate at all Question 4 To what extent have you experienced symptoms health complaints during the past week O No symptoms health complaints O To a mild extent O To a moderate extent O To a severe extent Figure 2 7 Injury questionnaire 4 2 3 Participatory sensing Participatory sensing is a concept about use the mobile devices to form interactive participatory sensor networks that enable public and professional users to gather analyze and share local knowledge 22 This concept moves the data sensing from observing devices to the users An example of participatory sensing architecture is shown in Figure 2 8 The presented figure shows the architecture in participatory sensing with use of mobile devices to capture data then sent to a server and presented afterwards From the previous section we mentioned how subjective data might give more precise data than objective data from third party devices because we can report how we actual feel compared to what the monitoring devices think we feel By following the participatory sensing concept we move the sensing part from sensor devices to mobile devices and the users are in control of what they wants to report According to a report from MediaNorge
45. asier to maintain 6 Hosting source code on a distributed server enables developers to continuously update the application without going through submission and approval to an application store Another advantage of using this approach is the support for platform independent the only requirement is a web browser Disadvantage with using this approach is when the device need to work offline without connection Without connection to the server the user cannot interact with the application or get presented any content because the content it s stored on the server Another disadvantage of creating a web application it the native API and other features the phone is providing is not accessible through a web application Hybrid application Hybrid mobile application combines both native development and web technologies 6 Hybrid applications look very similar to a web application because it is using web technologies to present the content but the application 1s wrapped inside a native container By doing this the application can use all native features the phone is offering To use this approach saves development time and maintenance cost because the source code can be reused on multiple platforms For example by using a hybrid application framework such as Cordova which is an open source framework Cordova provides JavaScript interface to access native features on the phone Disadvantage of developing hybrid application is that it cannot be d
46. athlete s wellness load and overuse injury based on the collected data from the players We are going to present the technologies and frameworks we used to implement a such portal And then talk about the push message service we created for the coaches to send real time messages to the players We end the chapter with a summarize 5 1 Motivation A mobile phone have limited screen to display content and power dependent Thus there is limited how much content we can present and what is optimal to present on a mobile phone Already players can monitor their own progression on load wellness and injury in application Have the coach monitor and analyze data through a mobile phone is not optimal From Section 3 4 we presented an Ohmage web portal and concluded that the portal was not created for the coach to use as an analyzing and a monitoring tool In the first version of pmSys coaches exported the data from Ohmage portal and then structured and edited the data manually in Excel From here they made calculations and presented the result in graphical charts with Excel We want to replace and ease this method of analyzing and processing data by implementing a web portal that presents the data captured from players parsed and calculated and then plotted onto graphical charts And have the coach focus on analyze and provide feedback to players From here the coaches or the medical staff can detect suspicious reports from their players and the
47. ation We then present two user studies we performed to support the evaluation of the new application In chapter 5 we present a implemented web portal for the coaches where they can analyze and monitor players based on reported data presented and visualized with graphical charts This web portal is also able to remotely send push messages directly to the players mobile phone In chapter 6 we present a distributed sever which represent the persistent layer of pmSys Here we investigated the usefulness of deploying the system in various cloud services and studied improvements and optimizations towards this layer The three latter chapters are presented in parallel as shown in Figure 1 1 We summarize and conclude in the last chapter 1 Introduction 2 Background Related work 3 A player monitoring system using Ohmage 4 5 6 The pmSys mobile application The pmSys The pmSys web portal backend server 7 Conclusion Figure 1 1 Outline Chapter 2 Background and related work In this chapter we present system requirements for pmSys Then we are going to present sports related background such as what data we collect and why we are collecting it and how we can present it Furthermore we talk about a concept Participatory Sensing which is a concept where the participants do data sensing Further we present a system for collecting data from third party monitoring devices open mHealth Furthermore
48. ayer is ready for the upcoming match or training session and further see if the load is affected by wellness 3 Wellness data can help players recover between match and training for example reduce training load 3 Session rate of perceived exertion Duration Day Training activity RPE min Load Sunday Cycle 100 km a 180 940 Monday Jog 5 km extensive stretch a 25 50 Tuesday Skate 6 X 10 min at AT pace 5 min rec ig 120 840 Explosive weights abs Vis 40 280 Wednesday Cycle 30 km a 60 180 Thursday Skate 10 X 3 min at 5 km pace 5 min rec 8 75 390 Friday Jog 5 km extensive stretch ze 25 50 Saturday Skate 20 X 1 min at tempo 2 min rec 8 75 390 Explosive weights abs sj 40 280 Weekly load 3400 Monotony X SD 1 26 Strain load X monotony 4284 a A schematic example of training diary 1 Competitions Load Br 7 4a 56 B3 ru TT H Bi a 106 112 it 126 Time days b A visualization example of weekly load over a period 1 Figure 2 4 RPE monitoring 1 More tired than FATIGUE Very fresh F Normal Always tired normal SLEEP QUALITY Very restful i Restless sleep GENERAL MUSCLE lt EE Normal Increase in _ SORENESS eee i soreness tightness ane STRESS LEVELS Very relaxed Normal Feeling stressed Highly stressed Less interested in Snappiness at team others amp or mates family and activities than usual co workers A generally good mood Highly annoyed Very positiv Tar ery positive mood
49. be connected to one or more classes and campaigns can have one or more classes Within a class we have two roles restricted and privileged Within a campaign it exist four different roles participant author analyst and supervisor These roles is explained more on Ohmage s wiki page 31 A user s role is defined by the class role and the campaign role In pmSys we are not using the Ohmage roles to the utmost Instead we have defined our own roles player coach and coordinator researcher and system admin System admins are the ones who have responsibility for the system they are the ones who are creating users and assign a player or a coach to a team and giving them the right privileges System admins also have the responsibility to upload campaigns and assign the correct team to the campaign The player role is just a normal user in 21 Ohmage and corresponds to the participant role in Ohmage This user only can read their own data and have restricted role within a class A coach is a user with privileged permission within a class Privileged users are able to read all players data However the coach cannot read data from other teams in the system The coordinator researcher has more like an analyst role They have the same privileges as the coach in addition they allowed being an analyst on multiple teams The coordinator role can only access data they have assigned to If the system not assigns a coordinator to any campaign then
50. but it contains extra tools to motivate the user to use the application The application can receive push messages from the coach And the application provides better feedback than Ohmage with visualized charts We present the new version of pmSys with our new mobile application see Figure 4 27 In the next chapter we present the web portal created for the coaches to use with pmSys and evaluate this web portal against the ohmage portal presented in Chapter 3 In addition we present our implemented push message service Units Units per Montt 55 Proceeds Sales 0 Sep 1 2014 Apr 1 2015 UTC v 0 a Apple store TOTAL INSTALLS BY USER ON APR 27 2015 N YOUR APP Android 4 4 57 14 ll Android 5 0 17 86 Android 4 3 10 71 MM Android 4 1 7 14 E Android 4 2 BH Android 5 1 b Google play store 3 57 3 57 Figure 4 26 Total downloads of pmSys Ohmage front end Q ohmage pmSys Mobile application Dashboard Campaigns ohmage Ohmage server OS and Android Figure 4 27 pmSys with new application 56 Chapter 5 The pmSys web portal From the previous chapter we got a detailed view on how we implemented a new pmSys application for data collecting In this chapter we are moving the focus from players to coaches We are going to present a web portal we implemented to replace existing portal from Ohmage This portal can be used to analyze and monitor an
51. bz8jw v sif roscip 2 Y sif 62ipsi 5 Y sif zxy00u u Y Y sif ai3la9 v sif zhkf16 g v v sif uztl6m a a sif 32hody Type in message to send You have forgot to do Wellness Figure 5 13 pmSys push messages When the coach sends a push message through pmSys web portal the portal sends a HTTP request to our push message service and then our service sends a request to APN or GCM based on which device was requested And then APN and GCM send a push messages to the requested devices The devices subscribe it selves with the new application by sending a subscribe request to the push message service after the user have logged in As we discussed from previous section a coach could schedule a push message for when he or she want the message to be sent We solved this by extend the push service to support cron jobs by creating 66 Automated pushmessages E Name Time Repeat Message Created by Remove Wellness Dagsform 09 00 Sunday Monday Tuesday Wednesday Thursday Friday Saturday Remember App wellness dagsform sif a0zq3z Injuryreport 13 15 Friday Remember Injuryreport Skaderapport i App sif a0zq3z Injury report 14 00 Friday Remember injury report for this week sif a0zq3z sRPE 21 15 Wednesday Saturday Please remember sRPE after Game sif a0zq3z sRPE training 18 00 Thursday Please remember sRPE after training sif a0zq3z sRPE training 13 00 Friday Please remember sRPE after training sif a0zq3z sRPE training team SIF 20 00 Sunday Please
52. cess data 1 srpe kennet time 2015 04 25 srpeTypel 0 srpeWorkload 0 srpeType2 NOT DISPLAYED srpeLength 90 F time z 2015 03 31 srpeTypel 0 srpeWorkload 1 srpeType2 NOT DISPLAYED srpeLength 90 T time 2015 03 24 srpeTypel 0 srpeWorkload 4 srpeType2 NOT_DISPLAYED srpeLength 93 hy Figure 6 8 Optimized response read response Chapter 7 Conclusion In this chapter we summarize and present our achievements in this thesis Then we present possible future work 7 1 Summary In this thesis we have presented an implemented monitoring system for collecting processing and presenting data The system has proved to be an easier approach towards monitoring football athletes by collecting subjective self report data The combination of mobile application to report data web portal to monitor and analyze collected data and a back end server to handle and process reported and present data have formed our Player Monitoring System The players do not need to report by paper or excel sheets anymore Thus they can self report at anytime from anywhere and report as many times as they want using their mobile application The coaches can focus on analyzing and interpret the visualized data through a web portal rather than use time on processing and organize reported data and use third part programs such as excel to visua
53. coach view the players Readiness to play load wellness and injuries The charts here are also interactive See Figure 5 12 5 3 5 Push messages for coaches It does not exist any features in Ohmage system that could provide the coach a view of who have not reported for a day PmSys web portal provides a statistics over last 14 days reports but that does not help the coach to reminder the players who have not reported any easier Thus in pmSys web portal we present a overview for who have reported and who have not In Figure 5 13 we see a overview of who have reported RPE and Wellness for the day From here the coach can mark all or the players who have not reported and then send marked players a custom message Have the coach send push messages manually every day to remind the players 1s impractical in long run We want to ease this process by implementing a way to schedule a message and then be sent on certain time This way the coaches can for example schedule a reminder to be sent every morning to remind the players to conduct wellness survey With this feature we are not only helping the players to report but we have relieved the coach from remembering to remind the players to report every day In Figure 5 14 we can see a screenshot of scheduled reminders and how to add a scheduled reminder 5 3 6 pmsSys Push message service Our push message service consist of a MySQL database to store username and device tokens UUID and above t
54. communication Since we have build wanted to build ExpressJS http expressjs com 3Nunjucks https mozilla github io nunjucks Request https github com request request Sredis http redis io 6 Amazon SNS http aws amazon com sns TParse https www parse com 8 Appoxee http www appoxee com 60 the portal with NodeJS we decided to implement the push message service in NodeJS as well A web service that works besides the web portal to handle push messages In NodeJS it exist two libraries we could use to send remote push messages to iOS and Android devices node apn and node gcm In order to send push messages messages we are also dependent on the mobile device s associated UUID Once we have the UUID we could send push message with the libraries mentioned above The NodeJS libraries provides an easy interface to send messages to the platform message service and we does not have to read the APN and GCM documentation to set up a push message service An example of the portal can be found in Figure 5 4 Ohmage server Web portal st o st Client Database Web service Node Web portal Web browser Client request Push server to send push APN GCM Web service Token Validating database requests Push message to mobile device Figure 5 4 pmSys with push message service 53 Implementation 5 3 1 Login Since we wanted our portal to be used with diffrent Ohmage back end
55. cond contribution to this thesis we have developed a mobile application that works on two different mobile platforms The application itself provides a fast approach to self reporting and is customized to the athlete s use and needs The mobile application provides graphical visualization feedback which give a player the possibility to monitor their own progression and captured data The application works on Android and iOS devices and has until May 2015 been downloaded over 190 times We have also performed a user study to confirm whether this mobile application was better application that existing one and whether self reporting through mobile application was a better approach than pen and paper The third contribution in this thesis we have created a web portal to give visualized feedback and work as an analyze tool in the web browser This portal was created for the coaches to analyze and monitor their players and ease and remove the process of manual data processing and analyzing In combination with the web portal and mobile application we have implemented our own communication service between the mobile phone and the web portal allowing a coach to send push messages directly to a players mobile phone The last contribution of this thesis we have studied deployment of the system in various cloud services We found the network latency on mobile broadband to be slightly difference in the cloud compared to self hosting We also improved the s
56. cvdop Msif rv231e Wi sif 32hody Mi sif un qw5 Misif 4nriwd Wi sif t7qif3 MM sif a02932 a Load bil SRPE Team sessions 7 R A 6 PTA f Y A 4 5 DAN AAN r A a 4 f A MN tf j a L 2 3 o RON a er y ae lt lt MA S e J Y p nn V PS gt SS 2 y x N 0 23 09 2015 31 09 201 94 2015 08 04 2015 12 04 2015 17 04 2015 22 04 2015 23 03 2015 31 03 201 94 2015 08 04 2015 12 04 2015 17 04 2015 22 04 2015 b Team average Figure 5 9 Team visualization of Load Injury visualization On injury visualization we have presented the severity score based on the four key questions from Section 2 2 See Figure 5 11 64 lul Wellness Average Mood Sleep Amount Sleep Quality Soreness Stress Fatigue Readiness 5 a PAS A A A A So 7 AA 4 5 p R A y T FONS A sal A p a P a R o NG A E 2 Q PA ST e ALA 3 er e 4 NG S DE le SL AER EN E SS 4 E ai Fg 3 5 Ss EZ AS fs lt a Ta SI gt ELAN EE rn ZS a za A ER LASA A SA VAN PZ 3 2 A SR LEDD S LZ gt ESA RS DAS KG N i we a WT N S N Ww ef N TV No Ja WW y Te e y SY 25 Y d NY gt p Y 23 03 2015 28 03 2015 02 04 2015 07 04 2015 12 04 2015 17 04 2015 22 04 2015 27 04 2015 02 05 2015 07 05 2015 23 03 2015 28 03 2015 02 04 2015 07 04 2015 12 04 2015 17 04 2015 22 04 2015 27 04 2015 02 05 2015 07 05 2015 Misif l15
57. cy Policy v2 0 0 Choose an event Individual session Enter duration in minutes Previous i mD a Available surveys 105 dh onmage MWF Aii Surveys b Single choice Android c Number prompt Android onmage MWF All Surveys Choose an event O Match Team session Individual session Next Prompt Copyright O 2015 UC Regents View Our Privacy Policy v2 0 0 d Single choice 105 Date and start time for the session Select Date 13 May 2015 Select Time 18 39 Next Prompt Copyright O 2015 UC Regents View Our Privacy Policy v2 0 0 e Time stamp GOS Figure 3 7 Surveys and prompts in pmSys 3 3 4 Upload queue One of the great features Ohmage applications were providing was the ability to report offline regardless of internet connection This allowed the players to conduct any surveys at any time If a user conducts a survey offline the reported survey will be placed in an upload queue in the application The response will be stored in this queue until the user has internet connection again and the user has to manually upload the responses If the user has internet connection Ohmage MWF uploads the response right after the user has finished However on Ohmage Standalone was the response placed in the queue regardless of internet connection this happen because Ohmage standalone was waiting for GPS location and since we do not want to provide our GPS location the response wil
58. d Readyness to play means how ready physically and mentally are you to play if there is a match today tonight Visualization On the visualization page you can view your latest rating of fatigue sleep soreness stress and mood You can also view your ratings for the last 30 days A thick red line represents the average of the five wellness parameters Very fresh Very restful Feeling great Very relaxed Very positiv mood More tired than Difficulty falling asleep Restless sleep soreness tightness Feeling stressed Less interested in Snappiness at team others and or activities mates family and Always tired Insomnia Very sore Highly stressed Highly annoyed irritable down C 3 Injury survey 101 Injury registration When Once a week on a fixed day Registration contents Part 1 Please answer the following questions as best as you can 1 Have you had any difficulties participating in normal training and competition due to injury illness or other health problems during the past week 2 To what extent have you reduced your training volume due to injury illness or other health problems during the past week 3 To what extent has injury illness or other health problems affected your performance during the past week 4 To what extent have you experienced symptoms health complaints during the past week Part 2 If you have experienced injuries illnesses you will continu
59. dr Barroso Geoff Mulligan Adrian Hooke and Chip Elliott Internet predictions IEEE Internet Computing 14 1 12 42 January 2010 6 Phyo Min Tun Choosing a mobile application development approach ASEAN Journal of Management amp Inniovation 1 1 69 74 2014 7 Michael S Mikowski and Josh C Powell Single page web applications JavaScript end to end Manning 2014 8 Prototype mobile applications built with ibm worklight for ibm watson http www ibm com developerworks library mo prototype watson Accessed 2015 05 13 9 Ajax new approach to web applications https web archive org web 20080702075113 http www adaptivepath com ideas essays archives 000385 php 10 Expressjs api http expressjs com 4x api html Accessed 2015 04 07 11 John Hicks Nithya Ramanathan Donnie Kim Mohamad Monibi Joshua Selsky Mark Hansen and Deborah Estrin Andwellness An open mobile system for activity and experience sampling In Wireless Health 2010 Academic and Research Conference 2010 12 Ohmage campaign prompt types https github com ohmage server wiki Campaign DefinitionfpromptTypes Accessed 2015 05 13 13 Belastningsoverv king i fotball http www nih no forskning prosjektarkivetl forskningsprosjekter ved nih belastningsovervaking i fotball Accessed 2014 10 12 103 104 14 Norges fotball forbund tippeligaen http www fotball no Landslag og toppfotball Toppfotball tippeliga
60. dy 74 74 0 sif dnrlwd B4 B4 D sif 62ipsi 128 128 0 sif 7ylan 147 147 0 sif a0zg3z 3 3 0 sif ai3la9 142 142 0 sif bgq7I0e g5 g5 0 sif eluewO 130 130 D sif h2dfbg 171 171 0 sif jxrokk 147 147 0 sif 11507gq 143 143 0 sif lwa z0 136 136 0 sif n9czhb 139 139 0 sif gl xgtj 129 129 o sif r chyg 119 119 o sif rosclp 109 109 o sif rv23 le 85 85 0 sif t7qif3 18 18 0 sif tdsw99 152 152 0 sif unbqws5 65 65 o sif uztlom 143 143 o sif zhkf16 121 121 0 sif zxyOou 147 147 o Total All Users 3133 3133 oO Figure 3 13 Leaderboard in web portal 3 4 6 Hard coded endpoints We analyzed the front end s source code and we found all the API endpoints were hard coded and pointed to the URL app Which means Ohmage backend server have to be hosted at the same server as the web portal and must be hosted under the URL path app Otherwise would the front end not work See Figure 3 16 It would be easier and more portable portal 1f we could pick the endpoint server our self in case we want to run the Ohmage server somewhere else For example we could host Ohmage back end server in the cloud and have the web portal hosted on a local server 3 4 7 Conclusion We conclude that the web portal worked as a user management tool because it provided an easy and clean user interface for user manipulation and uploading a campaign However except from these two features the coaches did not found the web portal very useful as a monitoring application for th
61. e sensors and the application 30 and we did not get this feature to work either The profile and help on both application was to customize application settings and present various application guidance 3 3 2 Campaigns and projects In Ohmage standalone was the campaigns feature called Projects and on iOS was called Campaigns In order to conduct a survey we need to download the campaigns we want to conduct In both applications were the approaches with download of campaigns the same We click on the feature and then we need 23 tt onmage MWE gt M la Saving screenshot Q ohmage Campaigns Surveys Queue 0 Projects Surveys Response History 2 E h ik Pm Profile Help Reminders Upload Profile Mobility Queue Copyright O 2015 UC Regents View Our Privacy Policy v2 0 0 a Sy Streams Help gn CI al a Dasbhoard screen Ohmage MWF b Dashboard screen Ohmage MWF 105 Android Figure 3 5 Screenshot of Ohmage dashboard screens to add available campaigns before we can conduct the surveys Both approaches are briefly presented in Figure 3 6 o my Projects You are not participating in any projects Available Surveys Available Campaigns ee i No Available Surveys Demo Survey2 Players Se Please install a campaign to view gt available surveys SIF Survey Players e Upload Queue Refresh Campaigns Copyright O 2015 UC Regents Copyright O 2015 UC Regents View Our Privacy Policy View Our Priv
62. e application stores the response into a file This approach makes the application only require internet connection the first time we log in and right after login we never have to log in again because all necessary information 1s stored on the file system Our implemented file system is also user based which means every time we have logged to the application with a valid username and password The application creates a folder that corresponds to the username From this point every files created belongs to the logged in user and is stored within the user s folder If the user decides to log out all stored data that belongs to the user will be deleted This allows the application to have multiple user application in the future but have not been implemented in this thesis E kennet gt EX login gt server http lastLoggedin txt pa program j ohmage nktconsulting net B login gt B visualization ie app username kennet password 2a 125tDUDdLSbpR hgBOBVunej O3iNpxX6s FK Uca rFinZgrTxgcW081ii cl ient pmsys app teamId demo Figure 4 16 Example of pmSys file system 4 3 8 Offline reporting Reporting offline in Ohmage was possible and it is possible in pmSys to but we have done it slightly different Since we have downloaded all campaigns the first time the user has logged in all campaign and survey information is available in the application without internet The reporting process and presentation of
63. e total times a user perform a validation 6 5 3 Improved data fetching We wanted to improve the way data is sent from the server In our web portal and mobile application we have done a lot of necessary parsing on responses from Ohmage such as organizing the responses and connect the correct prompts to the correct values and date s see example of the old response in Figure 6 7 In Figure 6 8 we have illustrated our new implemented response read Here we can see that the data 1s presented in a much cleaner structure and make the parsing and interpreting for a developer much easier 6 5 4 Conclusion We ran a network measure on our new implemented endpoints and the result is presented in Table 6 4 We can see that our new way validation of username and password only used 2 ms In addition the new improved response read reduced the amount of data transferred by over 50 The tests were performed on the same machine as the testing environment Endpoint Bytes 346 bytes Validate username and password 2 ms 280 bytes 2335 byte New improved read responses 1080 bytes Table 6 4 Average latency and bytes transferred with new endpoints 6 6 Summary We conclude that hosting in cloud gave lower latency on mobile network than wired network Reasons to pick cloud services over self hosting are possible cost efficient guaranteed up time scalability backup infrastructure transparency The only time we should consider self hosting is
64. e with these questions 1 Is the health problem an injury or illness 2 Select the area that best describes the injury illness 3 Please state the number of days over the past 7 day period that you have had to completely miss training or competition du to this problem 4 Is this the first time you have reported this injury 5 Have you reported the problem to the medical device 6 Do you have more injuries to report What kind of information do we get The injury registration systemize information about acute injury overuse injuries and health problems The registration may detect health problems and symptoms before it develops into an overuse injury lt also record small injuries illness that are often overseen in traditional injury registration Clarifications It is important that you register all your health problems every week even if you have registered the same problem before or if you are receiving treatment for it If you have several injuries illnesses within one week be sure to record all of them by going through the registration several times Record the most serious injury illness first Your team physician physiotherapist fitness coach will receive a message when you record and injury It is important to emphasize that this system does not replace your regular contact with the medical team Please continue to make direct contact with the team physician or physiotherapist when you need it Visualization The visualiza
65. e within a module we have only one service that is distributed between all controllers and other services all controllers are stored and defined in one file However we have used AngularJS s dependency injection to injecting every service into each controller With dependency injection enabled in AngularJS we can reuse the services over and over again Visualization Module Visualization Rpe module 3 visualization visualization html Y amp app visualization js b O about visualizationControllers js Controller View O common visualizationInjury html visualization Service 7 dashboard ES pa Pulling and fetching ep Visualization Wellness module visualizationMenu htmi data from remote server Containers 3 login B visualizationServices js CJ profile visualizationSrpe html Service l visualizationSrpeLoad html model Controller View SEN see visualizationSrpeRpe html O reminder visualizationSrpeWeekLoad html Visualization Injury module b 7 survey visualizationTeamSearch html E E 3 visualizationWellness html visualization O visualizationWellnessLatest htmi 5 app js visualizationWellnessPeriod html Controller View a Screenshot of the b Screenshot of the visualiza c Example of our MVC application structure application folders tion folder Figure 4 6 Module based development 4 3 1 States and routes Usually when building web application the content is presented based on what the user have reques
66. ection 4 2 3 classic web application model synchronous user activity user activity user activity E E E E 13 4 8 2 3 S user interface system processing system processing JavaScript call HTML CSS data user interface Ajax engine Ajax web application model asynchronous HTTP request HTTP request user activity HTML CSS data XML data web server web and or XML server dient side processing y A y A ig S S 5 8 2 8 J S E datastores backend datastores backend z ig ls l a l processing legacy systems processing legacy systems E 13 5 13 E es Ez Ba Ba eee Classic Ajax h server side server side server side server side web application model web application model processing processing processing processing Jesse James Garrett adaptivepath com Jesse James Garrett adaptivepath com a Ajax web application b Synchronous vs Asynchronous Figure 4 5 Ajax programming model 9 AngularJS The UI framework we mentioned in Section 4 2 5 Ionic is built with AngularJS and depends on AngularJS to function AngularJS is Open source software and a JavaScript framework maintained by Google Some of the features AngularJS is providing 39 Two way binding gives automatic synchronization between the view HTML and the model Data Structure for front end code by dividing the code into services controllers and views Ro
67. ectly to the database without validating the user and validating the data from the user can damage the database With the application layer we can validate and define business logic before the data 1s inserted or fetched from the database Ohmage server provides a public API over HTTP protocol Using HTTP methods such as GET and POST to access and provide data from the server The response from the server is presented in JSON format In short Ohmage server provides a public web service over HTTP protocol using JSON 50 The web service in Ohmage is built with Java and running in a java servlet container within the Java Virtual Machine JVM Ohmage is recommended to run with Apache Tomcat 51 Apache Tomcat is an open source implementation of a java servlet container running in Java Virtual Machine Ohmage server application is dependent on Spring Framework for dependency injection and Java Database Connectivity JDBC to communicate with the database Illustration of Ohmage server can be found on Figure 6 1 Servlet 1 1 I Data Validation 1 Database Connector n a Ohmage module layers and how HTTP requests b Ohmage server architecture is handled Controller Figure 6 1 Figure of Ohmage web application 11 Ohmage provides several end points which offer various services Our system is only using four API categories in Ohmage class API campaign API Authentication API and Survey response API Most of the requ
68. eir players With the first version of pmSys the coaches exported the data from Ohmage then organized the data by themselves and then generated graphical charts to present and analyze the data afterwards With a new web portal that is created for analyzing and presenting data where the coach can monitor and evaluate the reported data We also wanted to investigate whether it was possible for the coach to communicate with a player directly from this new web portal 3 5 Summary Ohmage was a great system for data collecting with self report but was not optimized and created for monitoring sports athletes We decided to keep Ohmage server because of the stability and user system Ohmage provide and build new application around Ohmage backend server We found both Ohmage mobile applications great as reporting tools but again they could be better by providing feedback through visualization of reported data The application could also have a faster report process by removing unnecessary user interactions By giving a better feedback on reported data by visualizing we believe this approach can increase the motivation for the user to report but also help the user to monitor their own progression We want to replace both existing mobile applications for the players and create one optimized and customized reporting application the users 30 l srpeLength Class Urn Class Name Description Members 8 014 ga an urn class sif coac
69. en Accessed 2015 05 02 15 Alt om fotball tippeliga 2015 http www altomfotball no element do cmd tournament amp tournamentId 1 Accessed 2015 05 02 16 Statistikk breddefotball http www fotball no nff NFF nyheter 2014 Statistikk breddefotball Accessed 2015 03 18 17 Peter J Denning Douglas E Comer David Gries Michael C Mulder Allen Tucker A Joe Turner and Paul R Young Computing as a discipline 1989 18 Lovdata Lov om behandling av personopplysninger Justis og beredskapsdepartementet 2000 USIKKER P DENNE https lovdata no dokument NL lov 2000 04 14 31 Accessed 2015 05 09 19 Benjamin Clarsen Ola R nsen Grethe Myklebust Tonje Wale Fl renes and Roald Bahr1 The oslo sports trauma research center questionnaire on health problems a new approach to prospective monitoring of illness and injury in elite athletes DENNE M FIKSES 20 Blake D McLean Aaron J Coutts Vince Kelly Michael R McGuigan and Stuart J Cormack Neuromuscular endocrine and perceptual fatigue responses during different length between match microcycles in professional rugby league players International Journal of Sports Physiology and Performance 9 367 383 2010 21 Rated perceived exertion rpe scale cleveland clinic http my clevelandclinic org services heart prevention exercise rpe scale Accessed 2015 04 16 22 J Burke D Estrin M Hansen A Parker N Ramanathan S Reddy and m B
70. ensing platform Technical report UCLA Computer Science Auguest 2014 Ohmage front end wiki https wiki mobilizingcs org app web Accessed 2015 04 22 Ohmage http ohmage org Accessed 2015 03 07 Mwf mobile web framework university of california http mwf ucla edu Accessed 2015 03 07 Google play ohmage mwf https play google com store apps details id org ohmage mwoc hl no Accessed 2015 03 23 Cordova https cordova apache org Accesed 2015 03 27 Cordova plugins registry http plugins cordova io Accessed 2015 04 27 Glenn E Krasner and Stephen T Pope A cookbook for using the model view controller user interface paradigm in smalltalk 80 J Object Oriented Program 1 3 26 49 August 1988 Ionic documentation overview http ionicframework com docs overview Accessed 2015 03 28 Sandeep Panda AngularJS Novice to Ninja 2014 Angular ngtouch documentation https docs angularjs org api ngTouch Accessed 2015 03 29 X2js working with array https code google com p x2js wiki WorkingWithArrays Accessed 2015 03 31 Apple developer app life cycle https developer apple com library ios documentation iPhone Conceptual iPhoneOSProgrammingGuide TheAppLifeCycle TheAppLifeCycle html Accesed 2015 03 02 Apple push notification service https developer apple com library ios documentation NetworkingInternet Conceptual RemoteNotificationsPG Chapters ApplePushSer
71. er interface design how the applications look navigation How it was to navigate within the applications and how the content was presented The questions presented were applied for both application For each question they got five alternatives 1 Poor 2 Ok 3 Good 4 Very good 5 Excellent We ended the survey with a question on which application they prefer ohmage pmSys or both the application 1s equal In Figure 4 22 we can confirm that the majority found pmSys application better than Ohmage Most of the users found pmSys very good and excellent on the four fields we asked them But with Ohmage the majority found Ohmage Okey which 1s less great than Good 20 of 26 our testers said they prefer pmSys over ohmage See Figure 4 23 We could have invited more people to get a better result but due to our time limit we only reached out to 26 people We got overwhelming comments from our test users such as pmSys provided better questionnaires than Ohmage and it was easier to navigate through the application pmSys provided a summary page of the responses which Ohmage did not have pmSys was simpler to use than Ohmage and provided more advanced content compared to Ohmage Visualization was a great feature to monitor activity pmSys provided more modern design than Ohmage pmSys provided simple features which resulted gave better user experience 4 4 3 User study pmSys as reporting tool We did not reach out to the natio
72. er to finish but also 26 la Ed 9 gt 84 44 Saving screenshot al MM Q Response History o Q Response History Q Response All Projects gt All Projects gt i Planned sRPE SIF Survey Coaches eng All Surveys All Surveys Apr 20 2015 7 54 PM gt lt March 2015 gt lt 3 23 2015 gt CALENDAR MAP sRPE after session 3 23 15 Date and start time for the session amp Mor Tue Wed Thu Fri SIF Survey Players 4 38 PM April 20 2015 at 7 54pm fil El 4 a 6 sRPE after session 3 23 15 Choose type of sesssion 4 38 PM 8 El m m IE SIF Survey Players Strength training 1 IB 19 Enter duration in minutes 2 90 3 Enter planned RPE En 31 2 Easy March 4 Total 4 a Calender b List of responses c Detailed response Figure 3 10 Response history in Ohmage standalone have less clicks The number prompt could also be improved by for example provide the possibility to do intervals on increment or decrement of a number instead of increasing the number with 1 we could increase the number with intervals of 5 instead Both these improvements would speed up the reporting process but also provide fewer clicks for a user to finish a survey Reminder feature was an excellent feature that helped users to do their surveys In our case this feature was very important In order to monitor players we need the players report data frequently and in time With the first version of Ohmage one of the feedbacks we got
73. es A better approach would be to save data in a SQL database on the phone This way the data cannot be easily read by the public Another improvements we should have done is to encrypt the whole file system Making the stored information protected from public exposion Native application If we want to archive higher performance on for example calculations we should consider native application development Native development provides access to native API features rather than running 1t through a container Indicate finished surveys Wellness surveys are conducted daily and injury surveys are done once a week It would be nice 1f we can indicate which survey have been done for the week or the day This way the player no longer need to check whether he have conducted a survey for the day or not 7 3 2 Web portal General web portal The web portal now is created for coaches to monitor and analyze the players We want to make this portal more for general use where the players can log in and monitor their reported data Converting user id s All username s is encrypted with a randomized generated string Only the coaches know which username that is connected to which player in real life The coaches found it hard to read the encrypted username s in the web portal Improvement can be done by converting the encrypted user 1d s and show the real name only in the coach s browser This needs be done on client side without sending any informa
74. estionnaires and a web portal to analyze and present data and a persistent server to save data securely All Ohmage applications are built around the back end server which means anyone can easily extend the system with new application to fit their use without removing the core functionalities of existing applications and Ohmage platform In next chapter we present the first version of pmSys using Ohmage Chapter 3 A player monitoring system using Ohmage We start this chapter with a discussion and evaluation from the previous chapter Next we presents pmSys using Ohmage and how Ohmage system fits our purpose and requirements Then we are going to present a mobile application we used for data collecting through self reporting and our evaluation around this application Furthermore we discuss and evaluate a web portal developed by Ohmage which 1s made for analyzing and presenting captured data We summarize with cons and pros with Ohmage which led us to new requirements and features for pmSys 3 1 System discussion The goal of this thesis was to implement a monitoring system to capture an athlete s load wellness and injury through questionnaires We wanted to modernize this approach by using modern technologies We presented the functional requirements for our envisioned system in the previous chapter in short the system must be able to perform three main tasks Capture Process and Present data We wanted to solve these tasks by part
75. eveloped independently of the OS and needs to be wrapped in or used with a native container The native container gives access to native features but the latest user interface might not be fully supported The native performance cannot be achieved with a hybrid application since it is wrapped in a container and uses the phone s browser to render content Conclusion Due to our time limit we limit our support to only create application for Android and 10S devices We picked the Hybrid application approach because we want to follow the principle build once deploy anywhere Thus developing a hybrid application with web technologies gives the other users who do not have an 105 or an Android device to use the application from a web browser without mobile device features In addition 1f we went for native application development we have to learn the platform specific language for example java for Android phones and Objective C for 10S devices But that is not all we have to develop two application in parallel We did not go for web application approach because of the need of accessing the phones native API In order to receive instant messages sent from a coach and storing and report surveys offline the native API was needed Thus was hybrid application the most optimal approach for creating a new application The three development approaches are illustrated in Figure 4 2 36 Application Source Code Application Source Code L Web
76. evices such as Apple iPhone Ohmage MWE is the 10S version of Ohmage and is developed by University of California UCLA 32 MWF stands for Mobile Web Framework and is a development framework by UCLA 33 The same Ohmage MWE application exists 22 on Android but has not been updated since 2013 34 Therefore Ohmage MWF version on android had various bugs that did not existed in 108 version The problem with Ohmage MWF on Android was that we could not change endpoint server we wanted to communicate with but this was possible on the 1OS version see Figure 3 4 b Android users were addressed to use Ohmage standalone version because the Android MWF version did not work The Ohmage standalone is not built with MWF framework and is developed by Ohmage team 32 Ill ff ohmage MWF Q ohmage Please Login Please Login to Proceed a kA EE ff onmage MWE Username Username Username Password Password Password Server Secure Login https play ohmage org Public Show Password register an account Login Secure Login Copyright 2015 UC Regents View Our Privacy Policy v2 0 0 v2 16 9 lt al al a Login screen Ohmage MWF b Login screen Ohmage MWF An c Login screen Ohmage standalone 105 droid Figure 3 4 Not able to select server on Ohmage MWF Android version The first version of pmSys was distributed to our users with both Ohmage mobile applications The goal was
77. ey items the questions and options are defined 20 in prompts prompt contains information about how a question should be presented including the question and various options In Table 3 1 we see which prompt type Ohmage supports and a description on what each prompt can do We are only doing questionnaires with a simple question and single option value as answer therefore the only prompt types we need and use are single choice number and timestamp Single choice is a prompt type with a question and multiple options presented but only one can be selected Number prompt consists of a number with min max and default value here the players can for example register how long their session was Timestamp is used in the campaign for coaches where the coach can set a time and date for their planned sessions and is used to report expected planned RPE value a session timestamp A moment in time by a user down to the second granularity the timestamp includes both date and time A free form piece of text provided by the user photo A photo taken by the user generally at the time that the user was prompted video only Android video taken by the user at the time that the user was prompted audio only Android An audio recording taken by the user at the time that the user was prompted Table 3 1 Available prompt types in Ohmage 12 A prompt can also have conditions on whether to show the prompt or not This way we can lim
78. from the coaches was Players wanted to report data but even with the reminder feature they tend to forget Hence we got a new idea what if let the coach help to remind each players to do their survey Because the coach know who have reported and who have not Response history only existed in Ohmage standalone and we found the feature relevant and useful However 1t would be better if we could provide players more useful presentation and feedback of reported data Instead of presenting a response history we could present the reported data in graphical charts This way the players can follow and monitor their own progression and have a training diary built in the application We believe that providing useful feedback with visualizing can increase the motivation to use the application as more than a reporting tool Both applications have a queue where the finished responses were placed in when the application could not upload the responses to the server In addition when the application has internet connection the user must perform a manual upload of the responses to the server Ohmage standalone would not upload the responses regardless of internet connection if the GPS were not provided We found this process inconvenient and a suggestion to this process is to remove the queue and use technologies to detect whether a connection is present or not and then automatically upload the responses Since one of our requirements was not save any data tha
79. h s third party data collection approach and methodology is relevant for our system mHealth Architecture Stovepipe Open Patient Caregivers Patient Caregivers J Analysis visualization MT L feedback Re usable health A data and knowledge services Standardized eee personal data vaults and health specific data exchange protocols MAP Pa AT ANOS Mobile platforms Mobile platforms iPhone Android iPhone Android Feature Phones Feature Phones a Stove pipe vs open mHealth 24 pe ra aaa See FN gt 4 5 8 i Y UD mu f A AS lt gt r of se A lt lt gt se ree aa el Pa pan de a eon eee p A Optimize your app s Integrate clinically Make clinical sense of potential usable data your data b Open mHealth system 26 Figure 2 9 open mHealth system 2 5 District Health Information System District health information system 2 is an open source system developed by Health Information System program HISP HISP is an open and globally network established distributed developers spread over countries like India Vietnam Ireland Tanzania and Norway and the development is coordinated by University of Oslo UIO and Norad 27 The system is used in more than 30 countries in Africa Asia and Latin America and these countries are trying to adapt DHIS2 as their national wide Health information system HIS 27 DHIS2 is a tool for collecting
80. have you experienced symptoms health complaints during the past week To a mild extent Is the health problem referred to in the four questions above an injury or an illness Illness What is the illness related to Stomach region Please state the number of days over the past 7 day period that you have had to completely miss training or competition due to this problem 1 Is this the first time you have registered this problem Yes first time Have you reported the injury illness to the medical team No ave you experienced any other illnesses injuries or other health problems during the past 7 days No Back to program b Take part 2 Figure 4 13 Reporting injury c No part 2 In pmSys we are only using three prompt types single choice number and timestamp An example of all implemented prompts are presented in Figure 4 14 On single choice questions we have removed the next button that existed in Ohmage applications We can imagine a player conducting the surveys several times throughout the season the player will over time memorize the questions and placement 45 of the options Removing the next button provide less unnecessary interactions and faster reporting However the next button will automatically be enabled on timestamp and number prompts Number prompt has also been implemented differently from Ohmage when the number prompt is presented the number can be incremented and decremented with intervals For example say the
81. he database we have a web service layer which communicate over HTTP protocol The web service is built with NodeJS and we use the plugins we mentioned in design Section 5 2 3 These plugins ease the process of setting up connection to a Apple Push Notification APN service or Google Cloud messages GCM and provide an easy way to send push messages The only requirements to use these plugins are we must have signed certification from Apple and Google and these certifications can be requested from Apple and Google with a developer account 65 Player breakdown Player readiness how ready is the player for match Load 30 day period RPE Load 8 7 6 5 4 3 2 13 04 20 04 28 04 05 05 13 0 Todays analysis Weekly training load 3 week period Wellness Training load Fatigue Everything is fine Low load yesterday Sleep Everything is fine Warning Medium strain hing is fine Soreness Everything is fine Injuries Stress Danger Must see medical staff Fit for training Mood Danger Must see medical staff E Monotony 1 1 E Strain 2473 02 05 UVO M Load W Monotony W Strain Sleep Amount Everything is fine 29 04 Wellness 30 day period Injuries 13 04 20 04 27 04 04 05 12 05 17 04 24 04 01 05 08 0 Mi Fatigue E Sleep Quality SleepAmount Mi Soreness Mi Stress Mi Mood W Average Mi sif ai3lag Figure 5 12 Individual visualization Players Select all Wellness sRPE sif qixqtj Y Y sif 4nriwd sif 2
82. he period March to May 2015 And this is the numbers from two different teams with 15 20 players These numbers proves that our application 1s a better application customized for an athlete and motives users to report frequently with the new application Total responses of old version of pmSys with Ohmage versus the new version of pmSys with new mobile application is presented in Table 4 2 Based on the numbers from Apple store and Google Play store 1s our application downloaded over 166 times on 1OS and 28 times on Android Screenshot of total Apple and Android downloads in the period September to April are presented in Figure 4 26 54 RPE report Wellness report og LO ep pr OD OD C C O r O O O vu Z op no T 00 O Ko 00 pmsys Ohmage pmsys Ohmage a b Reporting non injury Reporting knee injury 00 N N gt v V e E O O O O N 10 N Oo pmsys Ohmage pmsys Ohmage c d Figure 4 25 Performance plot in seconds pmSys vs Ohmage Version Teams Total responses pmSys 2014 o 4 467 ohmage application pmSys 2013 gt 5005 new application Table 4 2 Total responses old pmSys vs new pmSys 45 Summary Based on user feedback from user study and the result of our application we have created an application that suits a sport athlete better than Ohmage The application is easy to use and is can easy be downloaded from the application store The application is only a reporting tool
83. hen process the captured data before we evaluate and analyze it A common approach to such research today is to use pen and paper to gather information For example observe and take notes with pen and paper or perform reporting through questionnaires Regardless the data must be collected processed and analyzed afterwards This process 1s not only time demanding but also impractical in a long run Norwegian school of sports science NIH 1s performing a research study about monitoring load and fatigue in Norwegian top football 13 Here they want to find the usefulness of subjective and objective tools to monitor football players in professional football Furthermore study the relation between a player s load fatigue physical performance and injury during a game but also after and before a game With help from these monitoring tools they want to see 1f 1t is possible to predict an injury or a reduced performance for a player The Norwegian top division Tippeligaen consists of 16 professional football teams 14 Each team has at least 20 players 15 which mean we have minimum 320 20 16 football players in Tippeligaen From a statistical report in 2012 by the Norwegian Football Federation NFF it is registered over 28 000 football teams in Norway 16 From a non technical perspective in football context pen and paper is still the primary tool and most used approach for data collecting Monitoring over 560 000 28000 20 football pla
84. her guidance 24 Today s health learning systems is often closed application and systems with their own proprietary data format management and analysis such as stove pipe architecture described in Figure 2 9 a Here we see that Stovepipe architecture 1s closed for reusing and sharing of data With open mHealth architecture we get an open interface for data capturing and reusable data for the users Open mHealth architecture is built around shared data standards and global communication networks With an open architecture components have well defined published interfaces which allow interconnection and use in ways other than originally implemented and intended 24 Open mHealth architecture is currently only capturing data from objective devices such as hearth monitor devices and more It is mainly focusing on saving and capturing health data for example body weight body height hearth rate and so on The captured data is used to provide visual feedback to be analyzed see Figure 2 9 b Our envisioned pmSys is about data collecting based on subjective data and with questionnaires Open mHealth does not provide methods to capture data through questionnaires and neither does the system provide any applications to present data Open mHealth is great concept and Read more about open mHealth here http www openmhealth org 13 system but not ready to be used as our system platform yet However as a part for future work open mHealt
85. hes 1 SIF Coaches a Response distribution tw s amp tame are ate tam Bee e Suda ny um y om at vann royo et mee Cote eee apma aso om e aa gt ewe o wept ie o srpeWorkload b Scatter plot e sweet we alas ay e ds at a e a e i gt e e a Figure 3 14 Visualization in web portal Make Privileged Make Restricted Remove L Add Members Username User Role nih havard wiig ohmage admin sif 4bt7s2 sif a0zq3z sif hbr192 sif iSfwaq sif ijflud JE Privileged JE Privileged JE Privileged JE Privileged JE Privileged JE Privileged JE Privileged a Coaches Class Urn Class Name urn class sif players 1 SIF Players Yo Description Members 32 Make Privileged Make Restricted Remove values Mo Rest 1 Very very easy 2 Easy 3 Moderate i Somewhat hard 5 Hard 6 7 Very hard me 9 10 Maximal Add Members Username User Role ohmage admin sif Ow6 16j sif 1396v0 sif 27py9l sif 2bz8jw sif 2cvdop sif 32hody JE Privileged Restricted Restricted Restricted Restricted Restricted Restricted b Players Figure 3 15 Class administration in web portal We also want to create a new portal for the coaches to work as a better analyzing and monitoring tool The Ohmage web portal only presented visualization on statics for the collected data but we could w
86. hmage server already handle this for us However with a portal built around Ohmage backend we get a portable and independent web portal that is not server dependent We can install and run the web portal anywhere we want and allow the user to pick which ohmage server they want to log onto If we look back at the Ohmage portal not only was the API endpoints hard coded to a specific path and url but the portal must also be hosted at the same server as Ohmage backend 5 2 1 HTTP Client Ohmage s web portal runs in a java servlet container and fetches and sending data to Ohmage server through HTTP requests We want to implement a such web application with NodeJS instead of a Java Node JS is a platform built on Chrome s JavaScript runtime for easily building fast scalable network applications It uses an event driven non block I O model that makes it lightweight and efficient perfect for data intensive real time application that run accross distributed devices 45 NodeJS is very easy to get started with and as we can see in Figure 5 1 with minimal lines of JavaScript code a HTTP server is created this example is taken from NodeJS website 46 NodeJS is very portable and can be installed on every desktop platforms Windows Linux or OSX NPM 47 is NodeJS s package management to handle library and package dependencies var http require http http createServer function req res res writeHead 200 Content Type te
87. id longer than 15 minutes Result 1s presented in Figure 6 6 select x from preference where p key auth token time p key p value auth token time 10800000 Hoooooo Figure 6 6 Auth token in database 6 5 2 Validating users When we created the pmSys mobile application and pmSys web portal we also implemented a push message service The problem with the push message service was 1t did not provide any form for validation when the user subscribe to the service We could not check if the user was an actual user 79 in Ohmage Ohmage does not provide a way to validate whether the username and password is valid or not The only to validate a user was to do a stateless authentication with the username and clear text password to the server But we did not want to store the password in clear text in the applications because we were afraid of possible exposures We extended the endpoint on the server to check whether the sent password was connected to the provided user The new extended endpoint provided significant lower latency compared to stateless authentication With the new validation endpoint the latency was low as 2 ms on the same server was the stateless authentication latency over 300 ms Providing a validating endpoint might leads to brute force attacks but considering security issues were not considered in this thesis However a suggestion to a solution would have been to limit th
88. ies 0 Rest 1 Very very easy 2 Easy 3 Moderate 4 Somewhat hard 5 Hard 6 7 Very hard 8 9 10 Maximal 98 C 2 Wellness survey Rating of wellness When Please rate your wellness every morning 7 days a week The rating must take place after getting out of bed but before training For example before or after breakfast or in the dressing room before the session Registration contents 1 Readiness to play Fatigue Sleep Quality Hours of Sleep General Muscle Soreness Stress Levels Mood OT O What kind of information do we get Wellness indicates how well the players overcome or responds to the training load and how well he recover A lower score than normal over time may indicate a higher risk of overuse injuries Clarifications Rate as best as you can according to the questions On the scale 3 is normal 1 er worst and 5 best Readiness to play has a scale from 1 10 where 1 is not ready at all and 10 is maximally ready Dictionary Fatigue means tiredness resulting from mental or physical exertion or illness Sleep quality means how was your sleep last night Hours of sleep how many hours did you sleep last night General muscle soreness means general soreness in the musculature especially in the legs Stress levels means a state of mental or emotional strain or tension resulting from adverse or demanding circumstances Mood means emotionally state of min
89. in 2014 81 percentage of Norwegian population have a smart phone device 23 Participatory sensing applications uses data from the mobile phone as a sensor to gather information in collaboration with its operator 22 A campaign model defines participatory application architecture This campaign model consists of four different roles Initiators who create campaigns and specify data collection challenges Gatherers the mobile users who participate in data gathering that may be network triggered user initiated or continuously reporting Evaluators who verify and classify collected data on behalf of the campaign Analysts the person who interprets and presents conclusions These four roles are taken from article Participatory Sensing 22 We wanted to build a system that is partly close to this concept where the reporting responsibility is moved to the users instead of to have coaches take notes and reports Participatory sensing concept emphasizes using various sensors located in the mobile phone to capture data However we have followed this concept partly by using the mobile phone to capture data with the user self report through 12 questionnaires Configuration User information Health wellness Personal data stream Location Proximity Activity Social context Spatial Context Media User personalization Device Device server Server Applications capture processing processing analysis action Sustainability
90. ing captured data How to Ohmage platform is built can be viewed on Figure 2 11 As a monitoring system Ohmage fulfill our purposes and requirements Ohmage have self reporting mobile applications on 10S and Android The ohmage platform also has a back end server with data storage and user management This back end is also open to be extended with new applications At last Ohmage comes with a web portal where the captured data is presented and visualized Web based Data Management and Administration Web based Data Visualization and Analysis sag e 0 Outer E IE row SPE 1 ohmage Web Project Survey Customized ohmage Interactive LifeStreams Passive Data FrontEnd Authoring Gathering Tools Explore Data Dashboards Visualization ohmage Backend Web APIs for access controlled data storage management and visualization p Q EN al SystemSens ohmage MWF Trialist ohmage AudioSens Mobility Self report Data Collection Apps Passive Data Collection Apps Mobile Data Collection Apps Figure 2 11 Ohmage architecture with applications built around Ohmage backend 2 8 Summary Our goal is to develop a system to monitor players through questionnaires and as we have already mentioned the combination of pen paper and excel sheet is not an option Online web services do provide a user friendly presentation of the survey but these surveys must be c
91. ion for example based on load wellness and injury but Ohmage provide statistical visualization In addition the portal can present the statistics of reported data the last 14 days Comparing this to Ohmage we have more clean and easier to understand feedback than Ohmage s total responses See Figure 5 17 userplot sif 2bz8jw Load 30 day period 00 EE o 0 a I o o o z o i q 00000 00 0000 om To VZ TERES 00 000 00 00 DE o 0 o ie oa a Pm 103 a RPE in Ohmage portal b RPE in pmSys portal Figure 5 16 Visualizing RPE in Ohmage and pmSys Response pmSys web portal is more focused on presenting the reported data and provide more useful feedback than Ohmage web portal Ohmage did not provide a way to view who have reported and who have not In pmSys we have presented a overview of who have reported and have not reported see Figure 5 18 68 count i i 1 i May 02 May 04 May 06 May 08 May 10 May 12 May 14 date a Statistics of responses in Ohmage portal Statistics of responses last 14 days 218 2418 225 20 o ca 18 O1 May 14 12 10 8 E sRPE 23 E Wellness 25 21 E injury 6 4 2 o 04 May 06 May 08 May 10 May 12 May 14 May E sRPE W Wellness W Injury b Statistics of responses in pmSys portal Figure 5 17 Visualizing total response last 14 days in Ohmage and pmSys Ohmage on the other hand let the user filter the responses th
92. ir research We knew that the system implemented from scratch would not meet NIH s wish thus we considered the idea of using existing systems We decided to try Ohmage as our monitoring system Ohmage provided applications we needed to conduct self reporting store reported data and a web portal for presenting data Ohmage have a server application that we could deploy to a distributed server The server application allowed us to define and create our own surveys and then present the surveys in Ohmage s mobile application With a distributed server approach the server can capture and process reported data at any time which provides the system availability In addition Ohmage has a web portal for users to view and analyze the reported data the web portal also provides basic visualizations At last Ohmage provides a user management system we could use to separate two teams in the system which means we can save data from multiple teams on the same server and Ohmage will provide data isolation and prevent unauthorized data access In Ohmage we were also able define our own username 17 18 Thus we could encrypt usernames with machine generated text and all user information and reported responses will be stored anonymously Only the coach with permission knew the connection between a user in pmSys and in reality This information was for security reasons not stored on the server or in Ohmage The first version of pmSys with Ohmage is presented
93. ired parameters to be sent within a request is username client name and hashed password or an authentication token The data is responded in JSON JavaScript notation If a requests fails the server respond with a detailed message on what went wrong The authentication API provides two types of authentication methods stateless and statefull Statefull authentication returns a token from the server which can be used to request other API endpoints Class API is used to get information about which class team a user belongs to and which privileges a user have For coaches who is privileged can view information about the whole class team which is a 2 A servlet is a Java programming language class that is used to extend the capabilities of servers that host applications accessed by means of a request response programming model http docs oracle com javaee 6 tutorial doc bnafe html Accessed 2015 05 11 3Spring http spring io 73 list of members of the class team For uploading and pulling surveys we use the Campaign API which provides information of about a campaign and the all surveys in the campaign The response object is presented in JSON format but the campaign and survey data is in XML format Clients who fetch campaign data from the server needs to interpret the XML formatted schema For response uploading we are using Ohmage s survey manipulation API here we can upload our responses and read uploaded responses A player
94. is filter is also supported in pmSys as well This feature was highly demanded thus a coach can contact the players who have not reported and tell them to do so 5 5 Summary In this chapter we have presented our implementation of pmSys web portal We have proved that our portal provides more useful features than Ohmage With features such as visualized feedback and this allows coaches to monitor and analyze their players The visualization 1s also heavily focused on a player s load wellness and injury rather than statistics We have increased the amount of reported data with our push message service and this feature was highly demanded from the first version of pmSys The portal have also made it easier for coaches to monitor and analyze their players in general We also got great feedback from NIH where they said pmSys was a very useful and good system to monitor the players and the system have helped their research a lot pmSys with the new portal and the application is illustrated in Figure 5 19 This web portal is not considered production ready even though the portal is used by the various coaches in reality But the portal have a great potential to be something big and to be extended with new features We can for example build a bidirectional communication channel between the coach and the player In next chapter we present the Ohmage server and our study hosting Ohmage server in the cloud We have also improved the back end server
95. it the questions presented and do not need to present them all The condition logic in Ohmage 1s based on which option the user picks on previous questions For example we can ask a player Was this a and the player can pick between 1 Match 2 Team session 3 Individual session Then the next questions 1s presented based on what the player picked previously If the player picked Team session the next question would be Which type of session with options like 1 Endurance training 2 strength training and 3 other However if the player picked Match from the previous question then the question about which session types would not be presented Possible prompt conditions in Ohmage are equal less than or equal lt greater than or equal gt not equal greater than gt and less than lt All these conditions can be used in conjunction with AND and OR Each prompt have a unique 1d within a survey for example idl for question one and 1d2 for question two The condition logic 1s based on these ids We can for example define a condition logic with id1 1 OR 1d2 0 for a question then the question will be presented based whether the idl was I or id2 was 0 otherwise the question will not be presented 3 2 3 User system Ohmage s user system is one of the central entities in ohmage and controls accessibility for the users It exists three entities within Ohmage users class and campaigns user can
96. ith a new portal present the data that is processed and analyzed before it is presented on graphical charts Furthermore the coach can focus on analyzing and interpreting data rather than processing data 31 public class AwConstants private final static String releaseServerLocation same as web server public enum AwUri x app is hardcodedx AUTHORIZATION app user auth token LOGOUT app user logout APP_CONFIG_READ app config read USER_CHANGE_PASSWORD app user change_password USER CREATE app user create USER DELETE app user delete USER INFO READ app user info read USER RESET PASSWORD app user reset password CAMPAIGN READ app campaign read SURVEY RESPONSE READ app survey response read CLASS READ app class read VISUALIZATION URL app viz WHO AM I app user whoami e And morass x Check https github com ohmage gwt front end blob master src edu ucla cens mobilize client AwConstants java Figure 3 16 Hardcoded Front end endpoints In addition we want to see whether it is possible to have a communication channel between coaches and players In the next three chapters present a new implemented pmSys mobile application first Secondly we present a new pmSys web portal created for the coaches At last we present optimizations towards the backend server and we have tested our s
97. l not be uploaded automatically thus the user must manually upload the response see Figure 3 8 3 3 5 Reminder Reminders are available in both Ohmage applications Reminder feature was used to schedule a reminder to help players to conduct surveys When the reminder was triggered it came as a notification on the mobile phone The reminder function in Ohmage standalone was hidden under campaign information and was not easy to find In additional Ohmage standalone also supports location based reminders which let us mark a location on a map or search for an address and then the reminder will trigger once 25 Mu la Saving screenshot ral ohmage MW All Campaigns fej Upload Queue All Projects gt sRPE after session Submitted on Mon Mar 23 2015 16 38 44 Planned sRPE 4 20 15 SIF Survey Coaches eng 8 00 PM I A Delete All Upload All This response is currently waiting for a GPS location before it is uploaded You can force the response to upload Dashboard now or continue to wait for the location Copyright O 2015 UC Regents Wait Upload View Our Privacy Policy v2 0 0 Upload All a One response in queue Android b Manual upload to server Android c View response before uploading Figure 3 8 Upload queue in Ohmage we get into our marked area see Figure 3 9 fe Project Info Create New Reminder SIF Survey Players urn campaign sif players 2 eng Reminder Survey Select a Survey to C
98. l panel They are saying that everyone can get a virtual machine up and running in as little as 55 seconds 60 Digital Ocean is offering more hosting locations than Azure and Amazon Digital Ocean provides hosting in Netherlands UK and Germany 61 Hardware specifications and prices is presented on Figure 6 2 a 6 2 b and 6 2 c We have only presented prices and hardware on the minimal requirement our hosting Our goal with testing Ohmage in cloud was to find which cloud services who could provide lowest latency and we tested the latency from Simula Research Labaratory and Telenors Mobile Network on 3G 6 4 Cloud hosting latency In our test environment in the cloud we ran Ubuntu 14 04 with tomcat 7 and MySQL server version 5 5 On amazon we ran t2 micro instance with I core CPU and 1 GB ram located in Ireland Azure ran on A1 Standard tier with 1 core CPU and 1 75 GB RAM in Ireland And Digital Ocean ran on 10 month with I GB memory and 1 core CPU We measured the average latency and compared the results with the Ohmage server instance hosted at University of Troms Amazon Azure and Digital Ocean We used Apache JMeter as our testing tool and we performed the latency test with 10 concurrent connections and sent 100 requests for each connection thread we ran the test from Simula Research Laboratory And on the mobile network we ran less requests due to limited available mobile network resources We only sent 10 requests in one
99. lize collected data We have also presented a unidirectional communication channel where a coach can communicate with a player A coach can also schedule custom messages to be sent to a player at certain time At last we have analyzed and studied the possibility of deploying the server in cloud 7 2 Main Contributions The main contribution for our system we have implemented a working monitoring system that is in use by football teams in Norway We had four teams using the first version of pmSys but due to lack of missing features and self report a new system was needed With the new version of pmSys and five football teams using the new version we have increased the total response tremendously with the new system Our system got media s attention after the Norwegian national team started to use 1t In addition the mobile application has been downloaded over 190 times on iOS and Android Our system is in use daily and we have over 50 players reporting every day We have also studied the possibility to deploy our system in Cloud to provide users high up time We conclude that our system made it easier for a team to monitor a players load wellness and injury We have digitized the approach to collect data through questionnaires by using a mobile application to self report The submitted data uploads automatically to a distributed server where the data 1s processed and stored securely The server has removed a coach s need for processing and orga
100. ly follow the participatory sensing concept about using mobile phone as a sensing platform and have the users conduct surveys through self reporting with their mobile phones Thus we need to present the captured data in order to analyze and monitor afterwards Therefore we must have a place to save the reported data from the users With a distributed server to collect and process the reported data from the users we can have applications and tools built around it to form our envision system With an open interface API like DHIS2 and open mHealth make the system both scalable and extendable by new applications Although with all systems mentioned in previous chapter we started to design our envisioned system from scratch We wanted the system to have a database to store all captured data and with a single database layer alone would not work because we must validate and analyze reported data before we can store 1t into the database 1f not anyone would submit random reports and upload data as they want With an application layer above the database the application layer can then handle all incoming and outgoing data before communicating with the database layer Then we have to consider user management and data privacy and as we designed and evaluated this idea the larger and advanced the system became new problems and requirements evolved Since NIH wanted to start their research as soon as possible and wanted to try our envisioned system for the
101. mes the fundamental to training when high load are being taken 2 All mathematical equations to calculate the various numbers is presented in Figure 2 3 Load RPEvalue x duration 2 1 Monotony WeeklyLoad x standarddeviationo f weeklyload 2 2 Strain WeeklyLoad x Monotony 2 3 Figure 2 3 RPE related calculations 2 With the weekly and daily load strain and monotony for a player we can create a weekly exercise diary and reveal an overall weekly pattern of exercises 1 We can also present the weekly load in graphical charts and allow the coach to have a visual impression of the session and the player s perception of the session In additional the coach is able to add expected planned load for the team session and analyze how a player or the team perceived in terms of to the planned load An example of weekly training diary and visualization of this diary is presented in Figure 2 4 2 2 2 Wellness We capture wellness data through five topics fatigue sleep quality general muscle soreness stress level and mood These five wellness values are evaluated and rated by a scale from 1 to 5 An example of such wellness sheet is presented in Figure 2 5 Once a day should the data be captured and are preferred to be done in the morning to get the best realistic results and are preferred to be done in the morning to get the best realistic results In combination with wellness score and weekly load can the coach determine whether the pl
102. ming model to make the application feel native and communicate with a server In the next section we present how we used these frameworks and technologies to implement the new pmSys mobile application 4 3 Implementation With help of the various javascript frameworks mentioned above and the existence of cordova plugins we used all technologies combined to build our application With cordova wrapping our application in native container The main JavaScript framework in our application was AngularJS We then divided the file structure in the application into module based folders For each module in the folder we have one or more views a service and a controller service is corresponds to the model in MVC architecture and the service layer is responsible for pulling data from the server and store it locally in a JavaScript containers We made a module based mobile application by dividing each feature into separate modules and folders In Figure 4 6 we can see that each module feature is represented by a folder and within each 3XML2JSON https code google com p x23s StringJSht tp if String S com SUnderscoreJS http underscorejs org 6MomentJS http moment js com INVD3 http nvd3 org 8D3 http d33js org AngularUI Routerhttps github com angular ui ui router 40 folder have many different view files The reason for it does not exist as many controller files and service files as view files is becaus
103. mo players 5 eng lt campaignUrn gt lt campaignName gt Demo Survey Players lt campaignName gt lt surveys gt lt survey gt lt id gt srpe lt id gt lt title gt sRPE after session lt title gt lt description gt Survey after session lt description gt lt contentList gt lt prompt gt lt id gt srpeWorkload lt id gt lt displayLabel gt Workload lt displayLabel gt lt promptText gt How was your session today lt promptText gt lt promptType gt single choice lt promptType gt lt Properties tag gt lt value gt 0 lt v lue gt lt label gt 0 Rest lt label gt lt value gt 1 lt value gt lt label gt 1 Very very easy lt label gt lt 1 More rpe options gt lt value gt 10 lt value gt lt label gt 10 Maximal lt label gt lt skippable gt false lt skippable gt lt prompt gt lt Other prompts here gt lt contentList gt lt survey gt lt survey gt lt id gt wellness lt id gt lt Wellness survey gt lt survey gt lt survey gt lt id gt overuse lt id gt lt l Injury illness survey gt lt survey gt lt surveys gt lt campaign gt Figure 3 2 A tiny example of a Campaign file 3 2 2 Surveys Within a campaign it contains one survey collection and each surveys within the collection is identified with a unique name We have three survey items within the collection and the three survey items consist of RPE wellness and overuse injury Within all the surv
104. n Third part devices pmSys would be a better monitoring system 1f we could integrate third party monitoring devices and present the data The coach can then compare subjective data versus objective data Other database technologies We can study other RBDMS such as PostgreSQL or noSQL database such as MongoDB and see if we can find any improvements for the server by replacing the MySQL database Security issues Study and improve security weakness of Ohmage server Update responses As for now 1t does not exist an easy way to update values for all responses For example 1f the RPE scale was distributed with a number from 1 to 7 Then suddenly the coach want to change the RPE scale from 1 to 5 If we want to do this with Ohmage we have to first update RPE scale defined in the campaign and then we have to iterate through all saved responses and update all RPE values and map the 1 7 scale to 1 5 scale 84 Appendix A Source code All pmSys repositories are private Access will be granted upon request A 1 pmSys app https bitbucket org nktteam pms app A 2 pmSys trainer https bitbucket org nktteam pms trainer A 3 pmSys push https bitbucket org nktteam pms pushserver A 4 Improved Ohmage server https github com knyyy server A 5 Ohmage server https github com ohmage server 85 86 Appendix B User Surveys B 1 pmSys vs Ohmage 87 Brukerundersokelse Fra en skala fra 1 til 5 helhets
105. n the staff can contact the player and provide further guidance We want to provide the coach a way to notify a player with a custom message while the coach analyze the players An example would be if the coach detects that the reported load to a player was to high and then the coach can contact the present player and make an appointment or just tell him to relax for few days By implementing this web portal we hope to ease the process of monitoring and analyzing for coaches and coordinators We want to replace Ohmage portal as a analyzing portal but keep Ohmage portal as a user management portal 5 2 Design This new web portal for analyzing and monitoring must work with Ohmage system in order to fetch various data from the players A web application that have directly access to the database fetching and process data through SQL queries could have been a solution But we do not want directly access to the database because Ohmage backend already have input validation and processing before the server stores the data the server also handles data access restrictions and user authentications We wanted to follow Ohmage s system model from Section 3 2 where applications are built around and communicating with the Ohmage server backend This way we also keep the database secured and 57 58 not manipulated or destroyed by unauthorized users Neither do we need to handle database transactions and follow the ACID properties because O
106. n minutes 90 How was your session today 1 Very very easy This is report for yesterday Figure 4 12 Summary page with option to report for yesterday told every player to report twice 1f they have more injuries or illness Among the injury question in the surveys we have a question where we ask Do you have more injuries you want to report Yes or No It would be better if we could restart the survey based on what the users pick here Ohmage application and back end have has not support for performing an action based on a conditional logic which means we can not tell the application to perform an action based on what the user choose Therefor this feature must be implemented in the application In pmSys mobile application we checks the finished response and look for whether it is an injury report or not and then we look for a prompt with id ouOtherInjuries And based on what the user picked yes or no the summary page present a button to take part two after submitting see Figure 4 13 Injuries lliness once a week Question 11 11 ave you experienced any other illnesses injuries or other health problems during the past 7 days a Injury question 4 3 5 Faster reporting gt gt Injuries lllness once a week Result success Have you had any difficulties participating in normal training and competition due to injury illness or other health problems during the past week Full participation without he
107. n our view Sscope message Everyone come and see how good I look y scotchApp controller aboutController function scope Sscope message Look I am an about page b scotchApp controller contactController function scope scope message Contact us JK This is just a demo Hz Figure 4 7 Example of routing in AngularJS var app angular module pmSys app config function stateProvider SurlRouterProvider SstateProvider State visualization I url isueli sation templateUrl app visualization visualization html controller VisvalizationController state visualization menu parent visualization params uern templateUrl app visualization visualizationMenu html controller VisualizationMenuController State visualization teamSearch parent visualization params urn templateUrl app visualization visualizationTeamSearch html controller VisualizationTeamSearchController state visualization wellness parent visvalisation params uern templateUrl app visualization visualizationWellness html controller VisualizationWellnessController state visualization wellness latest parent visualization wellness params vre templateUrl app visualization visualizationWellnessLatest html controller VisualizationWellnessLatestController Hz
108. nal team and Sarpsborg 08 due to ours and their time limit But L renskog IF and Str msgodset IL were positive to perform our user study to the system This user study was performed to get a confirmation on whether pmSys was a better reporting tool than other tools such as Excel Pen and paper or Web surveys And how they experienced the usability of pmSys mobile application We got in total 30 responses and we used the scale we discussed in previous section from I to 5 We asked how they felt about RPE Wellness and injury reporting process and which reporting tool they like to use Before using pmSys L renskog IF reported RPE and wellness through Excel sheets and sent the response over email The outcome was that everyone preferred pmSys over excel and pen and paper We also got confirmed that pmSys saves time compared to the other tools Thus the result of preference and time saving 1s not presented on the graph The result of our study 1s presented in Figure 4 24 here we asked them to rate the usability design navigation how he content 1s present usefulness of pmSys and benefits of using pmSys We got some great feedback from our testers where some comment was e Great app Saves me a lot of time e pmSys is a great application to self report with e pmSys works phenomenally for me and I m pleased with how it has affected my life Positively e pmSys help me to track my training sleep and injury Based on the result presented a
109. nd the great comments we conclude that our application is a great reporting tool for sports athletes We did not get many participatory to our study due to our time limit but we are pleased with the result and can confirm that our application is a great reporting tool for sports athletes 52 4 4 4 Results Responses Responses Evaluation of pmSys Poor Ok Evaluation of Ohmage Poor Figure 4 22 Result of user study Ohmage vs pmSys Ohmage or pmSys Good a Good b Very good Very good pmSys Ohmage Excellent Excellent MA Usability MN User interface I Navigation MA Content MA Usability MN User interface I Navigation MA Content Figure 4 23 Application preference among our test users We wanted to perform a benchmark test to find which application could provide the fastest reporting process in both Ohmage and pmSys We reported reported each survey ten times on both Ohmage and pmSys and executed the survey as fastest as we could The reason we removed the next button was throughout a football season a football player will get used to the questions and answers presented in the 53 Rating of pmSys 16 MN Poor E o ES Good 12 HN Very good o MI Excellent G a B LR E Cr JE x Figure 4 24 User study pmSys ratings application and over time the player will memorize and create their own pattern for reporting knowing where e
110. nd visualized the data When the coach requested for visualization the portal sent a request to Ohmage server to fetch necessary data and when the portal got response from Ohmage server the portal sends the data unparsed clients web browser and let web browser render and present the HTML and charts We thought that having the client render the charts could relief the portal from server side rendering But we were wrong and we decided to move the rendering and parsing over to server site fixed the problem We have NodeJS both fetch and render the content from Ohmage and then present it to the user turned out to remove the slowness of visualization rendering After we moved the client rendering to server side rendering the visualization was not a bottleneck in our portal anymore 5 42 Push message service Push message service have increased the total responses A huge reason for this was that pmSys application could receive push messages and the coach could help the players to do their survey but 2 node cron https github com ncb000gt node cron 67 also send push messages at any time Another factor that increased the data reporting through the web portal was the coaches able to schedule a push message Right after we released this feature with the coach able to add schedule reminders The amount of responses reported increased In figure 5 15 we can see the result of scheduling a push message increased the reporting significantly righ
111. nes have limited storage and can be full over time With use of network connection we can store data at distributed servers then limited storage capacity is not such challenge anymore But at some point the application might need to store data on the phone s storage when the phone does not have internet connection All presented challenges is taken from the article by Phyo Min Tun 6 Released year Android 2007 2008 Apple pie Banana Bread Cupcake Android 10S 2009 sant Language Java Objective C Swift Tools Android SDK Xcode Eclair Package format apk ipa Froyo 10S 3 2 b 2010 b Gingerbread 10S 4 0 2011 Honeycomb 105 4 3 Ice cream sandwich iOS 5 0 2012 Jelly bean 10S 6 0 2013 Kitkat 10S 7 0 a Figure 4 1 Mobile operating systems release 6 Native mobile application Native mobile applications are installed directly on the mobile phone and can be launched without any container or intermediary tool 6 Native applications can all access the phone s operating system API s and use built in services such as camera accelerometer or location services GPS However developing native applications require deep knowledge of the specific platform When building native applications we have access to native API and UI and that gives us the ability to use the benefits of device specific DE Integrated development environment 35 software and hardware One
112. ng and the features we used to define pmSys with Ohmage These features were also one of the main reasons we decided to build pmSys with Ohmage 3 2 1 Campaign From Participatory Sensing concept we mentioned a campaign model which is a group of users participating in a program this is called campaign in Ohmage For our monitoring system is campaign representing a program for a team campaign is used to define various surveys for a team In a campaign we can define different surveys and within a survey definition we can define multiple prompts A prompt is an information type within a survey it contains information on how questions should be present and various options connected the questions The prompt type is whether a single choice question number question or time stamp we discuss these three prompts later A campaign in Ohmage is defined in XML format and contains a name and a uniform resource name URN of the campalgn In Ohmage the URN to a campaign have an important role we are using the URN to identify a team in conjunction with classes We come back to classes later A standard Ohmage URN consists of 19 urn campaign And in our system we are defining the urn by concatenating urn campaign with a team s initials and who the campaign is created for this 1s either players or coaches Then we end the campaign URN with a version number of the campaign and the language of the campaign This version number is updated e
113. nizing collected data from players The coaches no longer need to interpret collected data through pen and paper anymore The system automatically organizes and processes the reported data The coach only needs to deal with the web portal where the coach can view reported data presented in graphical charts and focus on analyzing and monitoring The coach can also send instantly push messages right to the players We have made a Player Monitoring System pmSys 81 82 7 3 Future work In this section we present possible future work and further improvements for our system 7 3 1 Mobile application Chat system The mobile application is the core of pmSys In our application we wanted to implement a bidirectional communication channel between the coach and the player This way the player can respond to the coach s message We only want the communication to be Coach to player and not player to player Custom message prompt We wanted to let the player provide a custom message after the survey 1s done this is to describe reported data further Multiple language support Language support is highly demanded for our application The users were both English and Norwegian speaking Some words in the surveys were hard to understand To solve this problem we want to have a localization file where every word is translated and then presented to the user Encryption and localstorage In pmSys now we are storing all information and data in JSON fil
114. ns which are downloaded from the server are received in XML format In the previous section we mentioned a library that helped us convert XML to JavaScript notation X2 S One downside we have with this library was how it interpreted single elements in XML For every single element in XML it converted the single element into a single variable in JavaScript However we wanted all elements in XML to be converted to arrays in JavaScript independently from from the XML element As we can see on Figure 4 11 a single element is converted to an single variable but multiple elements get converted to arrays According to X2JS documentation we could specify which key we want to be converted to an array before 1t got parsed 41 We were not aware of this and we implemented our way to convert the single variable in JavaScript to an array element in JavaScript lt campaign gt lt campaignUrn gt urn campaign sif players 2 eng lt campaignUrn gt lt campaignName gt SIF Survey Players lt campaignName gt lt surveys gt lt survey gt G lt survey gt lt survey gt lt survey gt lt survey gt lt survey gt lt surveys gt lt campaign gt lt campaign gt lt campaignUrn urn campaign sif coaches 2 eng lt campaignUrn gt lt campaignName gt SIF Survey Coaches eng lt campaignName gt surveys gt lt survey gt lt survey gt lt surveys gt lt campaign gt X2JS v Object campaign
115. of the biggest disadvantage of building a native application 1s that the source code written for one platform cannot be used on another platform for example source code written for Android cannot be used on an 1OS platform Native application development pattern is also time and cost inefficiencies due to duplicate effort for development and maintenance Another disadvantage of building native application is every platform require a platform specific SDK and we are forced use their unique tools Native applications needs to be distributed through the platforms application store the advantage of this 1s that we do not need to distribute our application directly to our users by ourselves However in order to apply and distribute our application to a such store we need to go through a large amount of policies and rules before our application gets approved and can be uploaded to their store Web application It exist two forms of web applications one where the view is rendered on server side and then sent to user s web browser The other approach is by moving the server rendering to client side and the client only focused on pulling relevant data from a server afterwards Both approaches are using HTML CSS Javascript to build applications Either way of doing it the application is accessed through the mobile device s web browser An advantage of using this approach is that the web application is less complex inexpensive faster to build and e
116. ologica Targets Data Store thinking a The information cycle of DHIS2 includes data collecting b DHIS2 architecture 27 processing analyzing presenting interpreting and provides feedback 27 Figure 2 10 DHIS2 platform 2 6 Other sport systems Milan labs is a research center for the Italian football club A C Milan and has been doing research on health objectives since 2002 28 Milan labs primary goal is to study the health of the players and identify guidelines to get the best out of the Milan players Milan Lab s most used methodology is by following up on each player wellness and use information technology to optimize and present the results Seattle Sounders from Major League Soccer MLS in U S A is also monitoring their players through wellness and RPE Most of every professional clubs have their own way to monitor and follow up their own players This information on how they do it or what they use 1s not often shared among the public Tromsg IL 1s one of teams in the Norwegian top division that uses monitoring devices such as XYZ to monitor their players in training and match 29 As we mentioned in Open mHealth section these monitoring devices and systems 1s closed for extending with new applications The users are also bounded to use their software and nothing else http www zxy no 15 Athlete Monitoring is a system for monitoring athletes through questionnaires on load wellness illness fatigue
117. onducted through a web browser and does not provide offline reporting After all responses have been submitted through online web surveys the coordinator still have to manually organize and analyze the data afterwards but also use time to generate readable charts before they can start to evaluating the result http www athletemonitoring com http ohmage org 16 We think that having a distributed server with a persistent database partly solves our problem because then the data 1s available from everywhere in addition organized and structured using database technologies The server can then validate reported data before it stores captured data In additional to the server we need to have a reporting tool and analyzing tool With a mobile application that is installed directly on the phone we perform self reports at anytime from anywhere and replace the pen and paper approach With a system that automatically process and present the captured data will let the user have focus on analyze and evaluate collected data rather than manually organize and process captured data The combination of mobile application analyze application a distributed server solves and relief the previous monitoring approaches and introduces a new method and process for monitoring The only system we have presented and discussed until now that fulfilled most of ours requirements were Ohmage Ohmage has a mobile application to conduct self report through qu
118. ontinue ES Select Time 16 41 SIF players Create new reminder Suppression Window privacy private 0 hours ae i i status participatin Time Reminder u participating Recurrence number of days responses 15responses submitted ef Location Reminder reminders 1 reminder configured Preferences Exclude Weekends Cancel Save Copyright 2015 UC Regents View Our Privacy Policy v2 0 0 a Empty project list b Add reminder GOS c Reminder was hidden on Android Figure 3 9 Adding reminders in Ohmage 3 3 6 Response history Response history feature was not presented in Ohmage MWF but only in Ohmage standalone This feature allowed us to view our response history in detailed We can filter based on various surveys and dates and view each response in detailed see Figure 3 10 3 3 7 Evaluation of mobile application Both applications worked as they were intended to do which was to provide a self report application It was easy to execute conduct a survey but the process of need to add a campaign before we were able to conduct the survey could be removed Hence the application will be faster and get to the point with self report The procedure of executing a survey could also be improved we could remove unnecessary interactions and let the user finish the survey faster An improvement towards this procedure could have been to remove the next button in this way the single choice questions would be fast
119. ork sia gass der r d SDS SANAE EE SEG 69 Figure of Ohmage web application 11 12 ENE AS NN LTE AEE 15 Latency on vred nevn uscar dl ceja Fess F TT ee aria a Ge 76 Latency on mobile broadband network 0 0 0 0 0 22 kr arr ren T1 Auth token timer was hardcoded to 15 minutes o a ao a 78 Auth tokenn database s y cis pe lA Be a Resa 78 Example of response read in Ohmage e 80 Optimized response read response aar rer rn rn rn kr nr rar 80 List of Tables dd 4 1 4 2 S 3 2 6 1 6 2 6 3 6 4 Available prompt types in Ohmage 12 20 Performance test pmSys vs Ohmage vnr nrk rann arr ren 53 Total responses old pmSys vs new pmSys LL aar rn rn nrk kr ee ren 54 PS SPT MESSE SEN noo es hw Fe FG ge Ta keg dd S 66 PM SVS push message service wn a2 Haw KR SA KE FE GE da 66 Few tables from Ohmage database knr nr eos 12 Average latency on wired network 2 2 74 Average latency from Mobile network eee ee ee en 78 Average latency and bytes transferred with new endpoints 79 Acknowledgements I would like to thank my supervisor P l Halvorsen for your great advice ideas and input throughout this thesis Your overwhelming feedbacks encouragement and commitment have been truly inspiring and I am grateful for the opportunities you have given me to work on an exciting project
120. orts athletes might have non technical background thus the system must provide high usability through user friendly design and simple user interactions The goal of the system is to provide tools and applications that are easy to use All features and user interfaces design must be simple to understand and not require a manual or any guidance to use Performance In this thesis we have not focused on hardware performances and specifications but we rather priorities improved performance related to the users Such as the reporting process must be fast and effective thus the user will report and use the system more frequently Moreover removing unnecessary user interactions also result in a faster system The persistent layer should be able to store data and provide data instantly The system in general should be fast and reliable Scalability The system must be able store large amount of data such as responses surveys user information and so on Storing large amount of data over time and handle multiple concurrent user interactions dependent on various factors Thus the system must be able to scale with the growth of the system Availability The system must be available at any time and from everywhere For the self reporting process the user should not be dependent on internet connection and should be able to conduct a survey at anytime which mean the mobile phone must be available for conducting self reports at any time The persi
121. os See FR SRM Oe ae h 37 FIT CON Las ed ah vee oot Yo Be ROA HAO BE PE Se SE Se 39 AS IMplEMENTACION dad rd TEEN ORDA CER TA SOO 39 FL Sates and EQUIES ga i se ara Berane Yh e Ged er 40 432 Campaigns and surveys ooo a 42 233 Report 1orvesterda 4 sansen sr ir a a as SS ELGG ts 43 254 Too EPON e ss Aske SA RS ee S LS Mick koke SE 43 255 Faster TepOnini ee 4 tag se TGS ed LA TTS ea a A Ess 44 43 60 Eropresenauons 2 sanse ES ra LENS SS 45 AO Flen stores SPREDTE STE EEE SER EN 46 25 8 OMMNACTCDOMING va ve skare a Mee et oD eae a SAG Sko ed 46 JD REMME h r whe eg aoe amp A ee Eg ark BE NG le Ge AG 46 TE Push NN ER ES de s 47 AOL VASUANZANON Gk ook ek rd SANT A SG 48 ad 2 ASIOSSAbY ode ad Sek PLEN ee ew en eS kr 49 44 Evaluation amp discussion sendid a eo ecen anna aa ee a G 49 AAA Hybrid application sd ea a an ie A die we GS es 49 4 4 2 User study Ohmage Mobile Application vs pmSys Mobile Application 50 4 4 3 User study pmSys as reporting tool 0008 51 GE OE A A A A A A AA E 52 245 MCONCLUSION 2 6 2 95 A A PR EES AR 53 15 o se ect GRS Ser hace cs Gave BF eae ae ae ahs d Oe a Ge oh GSA A es es ee 54 The pmSys web portal 57 SL Mouse US AT IAEA A EG SJ Be Dave Lee A AI eg S SL TT Gea SE sr Pr Dane EG a A 58 522 WebapplicanOm x as bi ke AG SAGE Sad G amp iers 58 525 Push messa SELVICE varsel re A PR OS GE 59 5 5 lnplementaviom s se kode EAN ds SE kee EE 60 A
122. osest to Troms to get lowest possible latency which means European servers is preferred 6 3 1 Cloud providers Amazon web services AWS is a cloud services who offers various cloud services all the services can be found on Amazon s home page 56 We mainly focused on Amazon Elastic Compute Cloud EC2 4Networkworld accessed 2015 04 13 http www networkworld com article 2866950 cloud computing which cloud providers had the best uptime last year html Saccessed 2015 04 14 http www cbronline com news enterprise it it services 10 top cloud computing providers for 2014 4401618 74 Which is a realizable virtual computer environment customized through a web interface 57 Here we can pick any operating system we want and install any application we needed Amazon server 1s available in Ireland and UK 58 Microsoft Azure is another known cloud service that offer various of cloud services On the same level as Amazon we are only interested in their virtual machine hosted in their cloud environment Azure is offering most of the same features as Amazon With a virtual cloud machine we can pick between various operating systems and install applications as we want In Europe is Azure providing two hosting locations Ireland and Netherlands 59 The last cloud services we want to study is Digital Ocean Digital Ocean is not as known as Azure and Amazon They provide Solid state Drive SSD only Hard drives and have a simple contro
123. over the system will only be tested with the free tiers in the cloud services thus we have not evaluated hardware specifications performance 1 4 Research methods In this thesis we follow the methodology design described in the report Computing as a discipline by ACM Task Force 17 This involves the design implementation and evaluation of existing systems and pmSys Since the system is being used in real life scenario we have identified and defined requirements and missing features in close cooperation with the users who also provided us great feedback and discussions For our development process we tried to follow Scrum software development methodology with a twist to 1t Each sprint consisted of planning developing testing and ended with a summary We tried to have 1 2 week sprints where we implemented new features and solve various tasks We described each task with various use cases that explained a user s approach to different problems which led to new requirements and features for the system All development tasks and use cases were organized described and discussed at Trello Trello is a free and visual way to organize anything with anyone We marked 2Trello https trello com each task on Trello with different priorities and followed the plan with 1 2 week sprints to implement new features and solve tasks We have not followed any testing frameworks or methodologies and we have neither created any unit tests for the im
124. ovides a great system to be used on desktop computers within a web browser the system is also heavily focused on health and personal data Thus DHIS2 does not have a mobile application where patients can report data but DHIS2 provides a web API that can easily extends the system with new applications DHIS2 cover another area and their purpose is to replace health systems national wide our goal is to implement a monitoring system through questionnaires using DHIS2 as a monitoring system throws away the whole purpose and potential of DHIS2 But we partly want to take DHIS2 web API architecture see Figure 2 10 b and use it as a requirement for our envisioned system Thus we can easily extend the system with new customized applications to the users needs Information Systems Web Portal o Document Repository Reports Plans Guidelines FAQs Clients Mobiles Smart Phones Tablets The Information Cycle Essential data set Definitions Tools Quality Reporting Information culture Feedback Actions Operational Systems Logistics Registries EMRs Web API Resource Abstraction Layer NV Collect N Target populations Use Process Predefined Data Web Sites News Weather Health Views Discussion Forums Interpret Analyse en em SE Data View Engine Charts Maps Data Tables Present Indicators Tables Graphs Population Maps Epidemiological pidemi
125. plemented applications However we tested the applications against various use cases and requirements We got help from test users at NIH to test the applications We have also performed a user study for the implemented system to confirm whether the system was fulfilling the intentions of a monitoring system 1 5 Main Contributions In 2014 we had four professional clubs from the Norwegian top division who used our system to monitor their players These clubs were Str msgodset IF Rosenborg BK Viking FK and Troms IL The system we provided them at that time was to collect process analyze and present data with use of existing system From this we learned and identified new requirements that led us into the development of a new improved system We name this system Player Monitoring System pmSys The main contribution of this thesis we have implemented a working system that is used by football players The system provides a monitoring system through questionnaires and with data collective processing analyzing and presenting tools This new improved version of pmSys is in use by to two top division teams Str msgodset IF and Sarpsborg 08 In additional the system is in use by the Norwegian national team both senior and under 21 team With the Norwegian national team using our system also drew media s attention and was featured in NRK s newspaper 18 November 2014 The last team to use the system is L renskog IF under 16 The se
126. pload Instead of showing a summary text like Ohmage do we wanted to give the user more useful summary page by showing a summery of their response before uploading they upload the response the user change the answers if necessary We have already presented a screenshot of it in Figure 4 12 with Report for yesterday feature Q sRPE after session sRPE after session Enter duration in minutes a Question s left pmSys b Question s left Ohmage Enter duration in minutes Figure 4 15 Presenting how many question s left 46 4 3 7 Filesystem On every mobile phone and operating systems we have limited memory available Most mobile operating systems kills background applications as they want to make memory available for running applications Apple devices with 1OS kills background application to free memory for running applications 42 In order to keep the application return to the exact same state as we left it we need to store necessary data to device storage in order to start the application from the same state as we left In pmSys mobile application we have created all necessary files and folder right after the user has logged in to the application We are saving login information and which server the user has logged onto In additional we are saving information about campaigns to present the surveys without internet connection and after a user have conducted a survey offline without internet connection th
127. presented to a numerical value from zero to 25 and these four numbers summarized gives a severity score from zero to 100 for each overuse problem The severity score 1s used as an objective measure of the consequences of an overuse problem and can be used to monitor the progress of overuse problems 4 These data 1s used to define athletes injury status and see 1f they have potential to moderate or severe reductions in sports performance or participation 19 Weekly questionnaire Diagnosis registration Project coordinator Athlete Weekly report Individual follow up Olympic medical team a Methodlogy for monitoring 19 4 Key Questions No problem reported mr Time loss days Who knows about it Previously registered Any comments No further problems b Diagram of injury report 4 Figure 2 6 11 Question 1 Have you had any difficulties participating in normal training and competition due to injury illness or other health problems during the past week O Full participation without health problems O Full participation but with injury illness O Reduced participation due to injury illness O Cannot participate due to injury illness Question 2 To what extent have you reduced you training volume due to injury illness or other health problems during the past week O No reduction O To a minor extent O To a moderate extent O To a major extent O Cannot p
128. r We wanted our users to use our portal without need of authenticate all the time In order to fulfill this we need a session storage to cache all data from the users throughout a session Instead of saving everything in memory we have stored every user data for a session in a redis database Redis is a open source lightweight key value database 5 2 3 Push message service We wanted to provide a way for the coach to communicate with their players by sending a message to announce something or just to remind the players to conduct their survey solution we mentioned in Section 2 2 was to send an email We wanted to take advantage of the smart phone by sending remote push messages directly to the players phone It exists a lot of providers that offer various push message services some of the services are Amazon SNS Parse and Appoxee A disadvantage we found with these services we have to pay the services use them We also wanted the coaches to send messages from our portal and use the already existing user system in Ohmage With these services the portal is also dependent on the third part service Some of services offered API access so we could integrate their service into the portal But again these services come with a cost Thus we wanted to control and create our own service Our first goal with our push message service was to allow a coach to send push messages to their players which means a unidirectional one to many
129. rage of the whole session fig 1 Take into account periods of high intensity running and periods of standing still e 10 is the highest exertion that you can imagine Imagine pushing yourself running a 3000 m test without any break e 0 is equivalent to rest and should not be used in combination with training e It is exhausting to sprint having a high heart rate fast breathing but also to tackle jump and duel How exhausted you feel in your muscles and mentally is also a part of the RPE e A match is typically rated 6 7 8 but could also be higher or lower e A strength training session is typically 2 3 4 5 because of much pauses Fig 1 Example from a 90 minutes session The assessment should reflect an average of the whole session Imagine that you rate every 5th minute of the session and then calculate the average value Rating of Session RPE Visualization e Training load is the RPE score multiplied by the duration of the session High training load occurs either by high RPE score high duration or both e Weekly load is the average training load the last 7 days Monotony describes the variation in training load over the last 7 days High monotony means low variation in training load Strain is the average weekly load multiplied by the monotony High strain means that the weekly load is high combined with low variation in training load High strain means less time for recovery and is associated with overtraining or injur
130. reference among our test users 22 arr ek kr nr re ren 52 User study pmSys falnsgs ica ee ee A E TEASE SEG a e 53 Performance plot in seconds pmSys vs Ohmage rar 54 Total downloads of pmSys acas z aome paise O a knr ra ran 55 pmSys with new application er rv rv nrk an 5 Hello World code in NodeJS listening for HTTP connections 58 pmSys web portal System overview 2 22 vr vr rv kran 58 HelloWorld in ExpressJS 10 ra ren 59 pmSys with push message service rv rn rer ran 60 Login screenshot of pmSys portal r er eos 61 An overview of pmSys portal e 61 NIC WING ATOS PONSE hr a ot ee EE SEE ST SS Se ce ans 62 Mere Tips average g s Sek le dos FASE AE ea kode sk 63 Team visualizacion of LO dd avs save sd STE SE bl teke er st 63 Example of wellness visualization 2 64 Team visualization of Injury r verna kr kr ren 64 Individual visualization 2 a 4 bee Soe eK Fet Ga SG eee GS 65 PSN YS PUSILMESSITES Le EN FEKTER PEER EGG 65 Screenshot of scheduled messages vr rn nr rn nr 66 Increased reporting after scheduled push messages arr vr vr arr rare 67 Visualizing RPE in Ohmage and pmSys 2 0 00 eee ra 67 Visualizing total response last 14 days in Ohmage and pmSys 68 Responses in Ohmage and pmSys verna kr nr rar 69 PRSYS Witl DEW P
131. remember sRPE after individual training team SIF game sif a0zq3z Figure 5 14 Screenshot of scheduled messages Endpoint subscribe HTTP Method Description To subscribe a mobile device to pmSys push service must provide username device platform 10S Android and device UUID token unsubscribe POST To unsubscribe a mobile device from receiving push messages must provide username send POST To send a push message must provide a user list of receivers users userId GET To get device information for a user Username must be provided fuser feams teamID Table 5 1 pmSys push message service new HTTP endpoints The new endpoints can be found on Table 5 2 We implemented this feature by using a NodeJS plugin with name node cron which allows us to schedule any task we want in NodeJS Endpoint HTTP Method Description cron create POST Creates a scheduled event using cron job syntax returns a cron job id cron stop Stops a scheduled event need to send cron job id cron getAllCronJobs Returns a list of all scheduled cron jobs for a team Table 5 2 pmSys push message service 5 4 Evaluation amp discussion 5 4 1 Visualization bottleneck In the beginning with the web portal our vizualization feature worked excellent when 1t had small number of data points to present But when 1t had more data points to render 1t became slow The reason for the slow rendering was how we parsed rendered a
132. rk r bare n oval D rlig Ok Bra Veldig bra Utmerket 2 Hvilken rapporteringsverkt y foretrekker du Mark r bare n oval Penn og papir Excel Web survey Mobil applikasjon Annet 3 Hvis annet hvilke 4 Sparer pmSys deg for tid ved rapportering Er pmSys raskere enn andre verkt y du har brukt Mark r bare n oval Ja Nei Fra en skala 1 d rlig til 5 utmerket hvordan vil du rangere helhetsvurdering 5 Brukervennligheten til pmSys Mark r bare n oval gt D rlig Ok Bra Veldig bra gt Utmerket 6 Designet grensesnittet til pmSys Mark r bare n oval D rlig Ok Bra _ Veldig bra Utmerket 7 Navigasjonen i pmSys Hvordan er det man vrere i applikasjonen Mark r bare n oval D rlig gt Ok Bra Veldig bra gt Utmerket 8 Hvordan presenteres innholdet i pmSys Mark r bare n oval D rlig Ok Bra gt Veldig bra Utmerket 9 Nyttigheten av pmSys sine funksjoner P minnelser for rapportering Visualisering Mark r bare n oval D rlig gt Ok Bra Veldig bra gt Utmerket 10 Utbytte ved bruk av pmSys Har du f tt noe igjen av bruke pmSys Mark r bare n oval D rlig C Ok C Bra Veldig bra C Utmerket 11 Har du andre kommentarer til pmSys Forbedringer nskede funksjonalitet Ris Ros Drevet av Google Forms Appendix C Surveys in pmSys
133. rs have an 10S or Android device and we limited our scope to only support these devices Developing applications individually for each operating system requires learning developing and maintenance in two different languages Synchronizing the design and features on both devices its yet another challenge With a hybrid application we can create create and develop a mobile application and only need to deal with one language and one application that works on every platform Hybrid application development work is great but in the end all hybrid applications needs to be compiled to native code in order to be installed on various devices This make it challenging when the development comes to 1OS devices For Android devices does the SDK work on every operating systems whether it is windows linux or osx But when it comes to developing compiling and testing 10S devices it cannot be 50 es N Telenor gt 10 14 61 MM glossary pmsys net pmSys glossar y Dashboard Den RPE N ew Norsk English 6 Velv re Wellness amp Profile ans SI F Skaderegistrering Reporting injuries Program SIF K NE CMJ il Visualization fr Reminder amp Glossary 1 About amp Logout es P AR mi a pmSys meny b Opens up in phone browser Figure 4 21 Glossary in pmSys done in linux or windows OS To compile and create apple s application file ipa we are dependent on Xcode which is apple s SDK to develop iOS application
134. rveys are defined and the responses are stored because of the restricted privilege the players cannot read each others responses The coach class is assigned to both player and coach campaign the coach campaign is used to report planned RPE and is only shared between the coaches By setting the coach to privileged in players campaign allow the coach to read every reported responses from each player The same distribution and allocation counts for the other team It is important to notice that the teams cannot view each others data And then we have the analyst and researcher team on top and they have the privilege role within a class and is assigned to campaigns they are research for The illustration describes the whole user system in pmSys Analyze research team Users Team alpha Team beta Coaches Players Privileged urn class analyze coaches 1 Privileged i Privileged Class urn class alpha urn class alpha urn class beta urn class beta coaches 1 players 1 coaches 1 players 1 Campaign urn urn i urn campaign campaign campaign i campaign alpha alpha beta beta coaches i players coaches players 1 eng 1 eng 1 eng i 1 eng Surveys i Surveys Surveys i Surveys Responses Responses Responses i Responses Figure 3 3 pmSys user system 3 3 Ohmage mobile applications Along with Ohmage platform Ohmage have two mobile applications for self reporting one for Android devices and the other one for 10S d
135. s This means that even with hybrid application approach of building once develop everywhere we still need each operating system s SDK An windows or Linux developer needs apple s SDK on mac in order to compile and distribute their application on 105 platform When it comes to testing each application needed to be tested separately on each platform By following the hybrid application development approach we saved a lot of developing time and we only have to deal with one language but in the end each application need to be tested on both platforms When we found the application usable and ready to distribute to our users we let our beta testers NIH try the application before we release it to get feedback of things we did not think of In order to distribute our application to our users we needed to follow the operating system s policy and rules before we could release the application We released many versions of our application through Google Play and Apple Store Google play used 1 4 days to approve the application before 1t released the application but Apple used up to two weeks and more to accept our application Because of the application stores used time to accept the application we have to set deadline for developing new features and bug fixes to two weeks before actual release day The conclusion of using hybrid approach is that in the end we still need all SDK s to build and test on the devices and we need the SDK to build
136. s a web application in a Java Server servlet container In the next subsections we present the features Ohmage web portal provide 3 4 1 Campaign Before we can start reporting data the campaigns with various surveys need to be uploaded to the server Through the web portal we can upload the XML defined campaign and it will then be parsed and validated Only system administrators are allowed perform this task 3 4 2 Viewing responses Users can use view their submitted responses in this portal Only users with privileged class role coaches can read and delete the players responses through the portal A screenshot from the portal 28 1s presented in Figure 3 12 In this figure we can see a filter to the left where we can specify start end date and filter the responses based on users and surveys Campaign Found 3133 responses by all users SIF Survey Players ka Shared responses can be viewed by anyone in the campaign Private responses are visible only to the responder and campaign supervisors Survey expand collapse All EJ gxpand cotapse Participant Wed May 13 13 29 52 GMT 200 2015 sif tdsw99 sRPE after session Private All y B Choose an event Choose type of session Enter duration in minutes Privacy individual session Football session A All Start Date How was your session today 4 Somewhat hard End Date Only photo responses Wed May 13 13 05 43 GMT 200 2015 sif qlxqtj Wellness every morning Private Show Re
137. s about 1 2 Problem definition Monitoring individuals in sport context with pen and paper involves not only data collecting but also data processing and presenting Use of pen and paper is not only time demanding and impractical but make the user spend more time on processing data than analyzing and evaluating which the latter 1s the purpose of monitoring With modern technology and use of internet we can automate most of the monitoring tasks In addition ease the data collection process but also make it fast and effective In this thesis we therefore investigate the question whether we can design and develop a system that can effectively ease the process of data collecting processing analyzing and presenting for monitoring Main goal will be to design a digitized approach to self reporting and monitoring We envision and aim for a Player Monitoring System pmSys and we will evaluate the usefulness of existing systems for pmSys As last in this thesis we want to find effectiveness of deploying the system in cloud and compare cloud hosting towards self hosting Since NIH already has their method for collecting objective data Thus we want to focus on collect subjective data through self report and not through external objective measurement devices potential approach to this problem area is to use a mobile phone to collect and capture data and then provide a web portal to present and visualize captured data with for example graphical
138. s who have not conducted their survey and remind them By using two different mobile applications we have to learn two different applications but what if we could deal with only one application that looks and function the same on every device Due to our limited time we decided to develop an application that is platform independent and followed the principle build once deploy everywhere In the next sections we present all technologies frameworks and tools we used to develop the new application We also present the result of our new application compared with Ohmage applications 4 2 Design In this section we present technologies and design decisions made for our application We are going to talk about native application versus hybrid application Moreover we present the frameworks and tools that we have analyzed and used to implement the new application 4 2 1 Mobile application development In the mobile application development context 1t exists three categories for development native hybrid or web application Considering which one of these three to use is also a dilemma 6 In mobile application development it exist five common challenges Multiple standards In mobile application development it iss hard to only deal with one standard because of the existence of multiple standards and platforms For example on Android devices 33 34 even on the same platform Android various smart phones operate with different hardware
139. s your session today 6 sif uztlom sif 4nriwo b Player list c Response data Figure 5 7 Viewing a response 5 3 4 Visualization Ohmage front end offered some basic statistics visualization However calculating reported data such as load strain or monotony was not possible in Ohmage Coaches have to export the data and visualize and analyze in third part application We wanted to create interactive charts and visualize reported data that have been processed before presenting to coaches The pmSys web portal visualization provides two types visualization groups the whole team and a single player Both the visualization groups are based the three surveys RPE wellness and Injury All visualization charts are interactive see Figure 5 8 which allows the coach to view detailed data on a specific player RPE visualization We visualize all players reported RPE score and load We have also provided a team average score on load and RPE for the coach the planned RPE score is also presented in the charts See Figure 5 9 Wellness visualization On wellness visualization we can choose to visualize Average Mood Sleep Amount Sleep quality Soreness Stress Fatigue or Readiness These variables are visualized based on what the player have reported See Figure 5 10 63 Lil SRPE Team sessions 0 nm NS N a 31 09 201 94 2015 08 04 2015 12 04 2015 17 04 2015 22 04 2015 23 03 2015 31 03 201 94 2015 08 04 2015 1
140. ser creation and support the users 2 2 Sports background In the previous chapter we introduced how we wanted to implement a system for monitoring sports athletes Before we can present such system we need to discuss the background on what information we collect and why but also how we use this information to monitor sports athletes A methodology to monitor sports athletes is introduced and discussed in these four articles e The Oslo Sports Trauma Research Center questionnaire on health problems a new approach to prospective monitoring of illness and injury in elite athletes 19 e Development and validation of a new method of the registration of overuse injuries in sports injury epidemiology the Oslo sports Trauma Research Center OSTRC overuse Injury Questionnaire 4 e new approach to monitoring exercise training 1 e Neuromuscular Endocrine and perceptual fatigue responses during different length between match micro cycles in professional Rugby league players 20 Here they monitored an athlete s Rating of perceived Exertion RPE wellness and overuse injury through questionnaires and self report These three subjects are also the basic to our data collection and monitoring approach thus we only want to collect subjective data from the players 2 2 1 Rating of perceived exertion Rating of perceived exertion RPE is a common method to rate the intensity of an exercise 21 In football context this would be cap
141. ser interface and the buttons and menus in the application to feel native Ionic framework is a hybrid mobile application framework that is focused on building native feel mobile applications by having components look and feel native 38 The reason we picked Ionic to build our new application with was because most of our users have 10S mobile phones and lonic provided UI components that look very like native UI component and the component gave a native impression 4 2 6 JavaScript With HTML and CSS to present the look of the application JavaScript is used to handle interaction between the application and the user JavaScript can also handle communication from a client to a server by using Asynchronous JavaScript and XML AJAX programming model On traditional web application when the user clicks on a button the client web browser sends a request to the server web Bootstrap http getbootstrap com 38 server and then the user have to wait for a response and when the server 1s ready to respond the client get redirect to a new URL address By using AJAX programming model the user gets a feel that the page does not wait for a response or need to be refreshed redirected 9 and make the application feel native when running on a mobile phone An illustration of AJAX application model and how AJAX handle the communication between a client and a server 1s presented in Figure 4 5 this approach is similar to what we presented in S
142. servers we have provided the user an option to pick a server on login This allows different teams the possibility to host their own ohmage server In case their want own and control the reported data and then use our portal to analyze and monitor their players screenshot 1s provided in Figure 5 5 We have limited the access to only allow coaches to use the portal 5 3 2 Session storage We mentioned a redis database in the previous section The reason for why we used a redis database is because we can save users data throughout a session for example after logging in to the portal we can save the password and username onto redis database to be used later when user fetches data from Apple push notification module for NodeJS https github com argon node apn OGoogle cloud messages for Android module for NodeJS https github com ToothlessGear node gcm 1 Universally unique identifier 61 Please Sign In Username Password Y Uit onmage http ohmage cs uit no Nkt ohmage http ohmage nktconsulting net Figure 5 5 Login screenshot of pmSys portal the Ohmage server And then we do not need to ask for the user password all the time The idea of using a redis database to store session data 1s taken from Modulus 48 a NodeJS hosting platform The redis database is hosted on the same platform as the web portal and is not accessible from the outside A system overview of pmSys portal can be viewed on Figure 5 6 lil S
143. specifications such as RAM CPU Screen Size and Storage capacity The same problems exist on 105 devices as well Lack of analysis tool Mobile application development has limited support for automated testing in native mobile application development For example testing sensors such as location service or multi gesture touch There is need for better analysis tools for measuring and monitoring various metrics of applications and testing Development platforms divide For example on Android which is an open source mobile operating system it exist various Android versions Some manufactures modify the OS source code and that results in multiple standards which then result in limited portability from one android version to another However 10S and Windows platform have their source code closed for the public thus platform divide is not an issue In the other hand 10S and windows application development requires IDE skills and knowledge which is not required in Android Frequent version releases Every operating system is releasing new versions of their operating system frequency see Table 4 1 to fix bugs and secure their operating system This creates different challenges for the developers Developers must learn different programming languages for each platform and new API release for every new versions In additional does the developer have to keep up with new software development kit SDK updates Limited storage capacity Most smart pho
144. sponses Rate your readiness to play a full match Fatigue Sleep quality 7 3 Normal Enter total hours of sleep General muscle soreness Stress levels 9 3 Normal 4 Relaxed Mood 00 4 A generally good mood Figure 3 12 An example of viewing responses through the web portal 3 4 3 Export responses The portal also provides a possibility to export all saved responses The data is presented in a CSV The coaches and coordinators used this function frequently to generate graphical charts through external programs such as Excel to present the data but also to analyze and evaluate the data afterwards This CSV file was not created to be open in Excel and was very messy the coaches coordinators needed to organize and process this file manually afterwards This workflow of processing and generating charts manually was not convenient and was not any different from using pen and paper and then manipulate captured data in Excel afterwards 3 4 4 Visualization The visualization feature in Ohmage was based on statics visualization Ohmage provide five types of visualization Total responses Single variable and Multi variable Total responses is visualizing based on number of responses the users have uploaded to the server but also present a leader board for responses See Figure 3 13 The Single variable visualization presents graphical plots based on a single prompt over a time The last visualization type let us present two different
145. stent storage should provide user access and present data regardless of users location and server location Security and Privacy The data collected in the system must be stored securely and not be accessed by unauthorized users Due to strict rules and policies in Norway we have to apply for permission to store and process sensitive information Therefore we only store anonymous and generated user id s that cannot be used to identify a person in real life Only the person coach coordinator who have applied for permission to the person data act Personopplysningsloven 18 know the connection between the encrypted username and the person player Nobody else knows this connection and in case the server is exposed for hacking the data will not be abused The connection between a user in the system and the person in real life is managed offline by the responsible person User roles The system should have a clear distinguish between the user roles The system should describe and provide four different user roles The players who provide the system with data through self reported questionnaires The medical staff and coaches analyze and interpret the captured data and gives feedback and results to the players The project coordinators and researches which have the same role as the coach but have access to multiple teams At last we have the system administrators which is responsible for system administrative tasks such as maintaining u
146. t after the release of this feature Our push message service is not production ready even though it being use in production now When we implemented our push service we did not consider security aspects of the system Hence we are validating every HTTP header and request for a certain value to be defined This is to prevent unauthorized users to spam our service with push messages Our push message service 1s still open to be abused if the user know the endpoints and use the service outside the web portal Statistics of responses last 14 days 22 8 5 204 2 o a 184 164 147 124 104 84 6 44 k ua h i E o Moa 04 Mar 06 Mar 26 Feb 28 Feb 02 Mar E sRPE Mi Wellness W Injury Figure 5 15 Increased reporting after scheduled push messages 5 4 3 pmSys vs Ohmage frontend Visualization Our visualization feature in the web portal provides more useful feedback for monitoring and analyzing reported data than Ohmage web portal Our Web portal provides customized charts based on the three survey s the players are conducting and the numbers are calculated and processed before presenting In figure 5 16 we have visualized a player s RPE value in Ohmage web portal and pmSys web portal As we can see Ohmage only provide a statics on how many times a player have picked a value However we can visualize a players RPE score and load value over a week and more We provide more specific visualizat
147. t could be used to identify a person in reality we must have every user to turn off the GPS acquisition in Ohmage applications To have applications that automatically record GPS location if the user is not aware of it could seriously harm our system and at worst break the law We want an application that does not record any location data at any circumstances With the injury reporting process we presented in Section 2 2 We wanted a player to report more than once if they have more than one injury or illness to report This could not automatically be done in Ohmage applications the users have to manually start the survey over again if they wanted to report the second time Ohmage does not support action based condition logic This means we could for example define a specific condition to restart a survey based on the user s answer We solved this in our first version of pmSys with Ohmage by telling the user to report twice if the player has more than one problem to report By the end of each survey ohmage presents a summary page with a message we defined on campaign 21 upload see Figure 3 11 Instead of showing a summary message a better approach would have been to show a summary page of answers the user have picked Then the user can perform an extra check and go back and change 1f they needed to la Saving screenshot 8 E ohinage MW Q sRPE after session Survey Complete Survey Completed Done with sRPE after session
148. tation javaee7sdk readme 1957703 html System_Requirements Accessed 2015 04 24 Amazon web services products and services http aws amazon com products Accessed 2015 04 25 Amazon web services amazon ec2 http aws amazon com ec2 Accessed 2015 04 24 Amazon global infrastructure http aws amazon com about aws global infrastructure Accessed 2015 04 25 Microsoft azure regions http azure microsoft com nb no regions Accessed 2015 04 25 Digital ocean features https www digitalocean com features technology Accessed 2015 04 25 Digital ocean reliability https www digitalocean com features reliability Accessed 2015 04 25 Amazon pricing http aws amazon com ec2 pricing Accessed 2015 04 25 Microsoft azure virtual machines pricing http azure microsoft com en us pricing details virtual machines Accessed 2015 04 25 Digital ocean reliability https www digitalocean com pricing Accessed 2015 05 17
149. te less and cleaner code With ExpressJS we do not need to handle HTTP requests in detailed such as manage HTTP headers on how the packets should be structured Hence we can focus on the content and how the content should be presented In Figure 5 3 we see an example of ExpressJS server that does exactly the same as plain NodeJS server from Figure 5 1 var express require express var app express app get function req res res send hello world Hz app listen 3000 Figure 5 3 HelloWorld in ExpressJS 10 Another advantage we found with ExpressJS was that it provides routing support Without ExpressJS we must handle all incoming requests manually and check the request object for what content the client requested and respond to the request This is handled automatically in ExpressJS Yet another powerful feature in ExpressJS is that we can pick any HTML rendering library we want We decided to use Nunjucks gt which is a template engine for JavaScript to render HTML content Nunjucks have support for inheritance of views where a child view can inherit from a parent view Nunjucks also provides conditional rendering on the HTML pages we could for example send variables to the view and the HTML page will be rendered based on the value we sent In order fetch data from Ohmage server we need a HTTP client in NodeJS Request is a simplified HTTP client in NodeJS which we used as to fetch data from Ohmage serve
150. ted from the server However when the application is running inside the Cordova container on a mobile phone The resource is accessed as a file system There is no existent of an address bar where we can type the address we want to access All resources and files is resolved by the Cordova container using the mobile browser s web view The start page of a hybrid application in Cordova normally index html unless we specify something else If we want to navigate another file we have to create a button with a link to access the other file and the web view will then resolve the request The Cordova container works like a web server serving user content based on requested resource from the user We wanted the application to feel native and with Cordova handling the requests like a web server would not provide the user a native feeling of the application Angular has a built in module for routing and linking ngRoute This module allows us to display content based on which resource we want to see and with each view presented it also provides a controller and a service on requested resource Figure 4 7 example shows how easy it is to define routes and assign controllers and views to various routes When building large applications the URL s become very long and challenging to maintain Especially when we sends long queries through URLs With Angular Ul Router we can define child views and parent views Instead of working with URL s Angular U
151. the message box over a period A better approach is if we could send a message that is presented to the user and then removed itself after the user have read it Android and 105 devices can receive remote push message notifications through the platform s notification service 10S devices uses Apple s push notification service 43 and Android uses Google Cloud Message 44 Receiving push messages on Ohmage was not possible In order for the application to receive a notification the mobile device provide the mobile phone s UUID to a notification service We have implemented our own push message service which we will present in the next chapter when a user logs on pmSys mobile application the application automatically sends the phone UUID to the push message service and then the application and the user can receive push messages We will present this more detailed in the next chapter As for now it is only the coaches who can send a remote push message to a player This allow the coach to give players instant feedback or just to remind players to conduct surveys This communication between the coach and the player is only unidirectional from the coach to the player but as for future work we want to extend this to a bidirectional allowing the player to answer to the coach Screenshots of push messages sent from a coach is presented in Figure 4 18 48 es N Telenor OSOA E 11 29 Wednesday 1 April e N Telenor 11 30 amp
152. the reminders in memory to present content to the user And then we have to save the reminder information to a file incase the application get closed by the operating system At last the reminders need to be added into the native platform specific notification service to schedule a reminder When the notifications are triggered 1t will be shown on the lock screen or in a drop down message from the top of the screen 1t will be shown even when the application is not running An example of pmSys reminder is presented in Figure 4 17 Oen Telenor gt 64 ML rs Reminder Reminder Title ARE Name Repeat Time On off 4 Y 4 Message Remember to do RPE Tuesday 28 April RPE weekly 14 00 Y Tine 14 00 Mon Rp Repeat Remember to do RPE Wellness daily 08 00 x Notification O Back Save Delete Add Reminder Cancel all a Add reminder b List of reminders c Triggered reminder Figure 4 17 Reminder in pmSys 4 3 10 Push messages Some players used the reminder feature and other got reminded through co players or coaches in person Most of the players said that they wanted to do surveys but they tend to forget it they always needed someone to remind them From the research study mentioned in Section 2 2 players got reminded by the coordinator who sent an email to remind Another solution could have been sending a SMS to the player Both these solutions require the message to be manually deleted after receive in order to not fill
153. through a mobile application Since we wanted to capture data with the mobile phone and present the captured data afterwards the data then must be stored on a distributed server With a client server model we are dependent on internet connection Therefore we must provide self reports for the user independently from whether the network connection 1s present or not on the mobile phone Data processing After a user has completed a survey and submitted the responses through their mobile phone The system should then be able to receive any incoming data from the users at any time the system can process validate and analyze the received data The system must have a persistent layer to store organize and secure the captured data However capturing and storing data alone cannot form a monitoring system therefore the system must also present and provide data to the client applications Presentation and visualization When the system has collected and saved the data then the system must provide a presentation of collected data With a web portal that is accessible from everywhere we can for example present the reported data in visualized charts Thus the data is easier to analyze and evaluate The mobile application should also provide feedback based on reported data this allows the users to monitor themselves and motivate them use the system 2 1 2 Non functional requirements Usability Most of the users who are football players and sp
154. tion indicates a severity score of the injury illness to be used in research Each of the questions 1 4 scores 0 25 and the larger the sum is the larger the severity score is It is important to emphasize that only your team physician or physiotherapist can diagnose and decide how seriously your injury or illness is Bibliography 1 Foster Carl Florhaug JA Franklin J Gottschall L Hrovatin LA Parker S Doleshal P and Dodge C new approach to monitoring exercise training Journal of Strength and Conditioning Research 15 1 109 115 2011 2 Aaron J Coutts Karim Chamari Franco M Impellizzeri and Ermanno Rampinini Monitoring training in soccer Measuring and periodising training 2008 3 McLean BDI Kelly V Coutts AJ McGuigan MR and Cormack SJ Neuromuscular endocrine and perceptual fatigue responses during different length between match microcycles in professional rugby league players International Journal of Sports physiology and Performance 5 367 383 2010 4 Benjamin Clarsen Grethe Myklebust and Roald Bahr Development and validation of a new method for the registration of overuse injuries in sports injury epidemiology the oslo sports trauma research centre ostrc overuse injury questionnaire Br J sports med 2012 5 Deborah Estrin K Mani Chandy R Michael Young Larry Smarr Andrew Odlyzko David Clark Viviane Reding Toru Ishida Sharad Sharma Vinton G Cerf Urs H lzle Luiz An
155. tion to the server Upload campaign Would be great 1f we could upload a campaign through the web portal rather than through Ohmage s web portal Machine learning It is possible to extend the system with machine learning on collected data The machine can then detect when a player is in danger to be injured and tell the coach Artificial intelligent We can combine machine learning with Al to automatically detect injuries and for example tell the player to relax for few days And other Al related stuff Chat system We mentioned this feature in section above we wanted to let the coach chat with a specific player through the web portal and then have player answer through their mobile application Campaign creator It would be nice for the coach and coordinators 1f they could create and edit campaign and surveys directly in the web portal 83 Planned RPE The coaches must use the mobile application to add planned RPE a better approach would be to add planned RPE values directly from the web portal Visualization We could let the coach specify which chart and how the data should be presented from the web portal This way the coach can create their own graphical charts 7 3 3 Server Text style in campaign In the campaign it would be great if we could style text in the XML file For example lt Question gt Rate your lt b gt readiness lt b gt to play lt Question gt and then readiness will be presented as bold text in the applicatio
156. to start monitoring and capture data at as early as possible Then evaluate the application to find new requirements and missing features and then implement a new application that is customized and created for the sports context With two different applications we quickly faced different challenges with our users Both them and we have to learn how to use both applications if would be easier if we only could deal with one application In the next sections we present the features in both Ohmage mobile applications on 10S and Android 3 3 1 General features After we have logged in to the applications we are presented a dashboard with access to all features in both applications See Figure 3 5 On Android we have Projects Surveys Response history Upload Queue Profile Mobility Streams and Help and 10S we have access to Campaign Surveys Queue Profile Reminders and Help We can see here that both application did not provide the same features We did not get Mobility and Streams feature in Ohmage standalone to work But the mobility feature was created to capture data on whether you walked ran or drove with use of accelerometer Wi Fi and GPS on the phone 30 Mobility data capture could been a useful feature for our system in the future such as collecting data on how far a person have walked or ran for a day and then predict a person s load for a day Data streams were created to send passively and continuously captured data from mobil
157. tured after a training session or a match The RPE value is given by a category rating from 0 to 10 This RPE scale is based on Gunnar Borg s RPE scale 1 An example of RPE scale is presented in Figure 2 2 where typical follow up question is How was your session and the present athlete respond with a number A reason for collecting subjective RPE data over third part measurement devices is because the objective measure devices such as heart rate monitors cannot tell how an athlete actual felt about the intensity of the session 1 By capturing the athlete s RPE value a daily exercise score load can be calculated by multiplying the duration of training match session and the RPE value By capturing RPE scores for a week we can present and calculate a weekly load for an athlete From this we can create training session indices such as monotony and strain and potentially provide an index of the likelihood Modification of the Category Ratio Rating of Perceived Exertion Scale Descriptor Rest Very Very Easy Easy Moderate Somewhat Hard Hard Very Hard Maximal Figure 2 2 RPE scale 1 of untoward training outcomes 1 Monotony is measure of day to day training related to the start of overtraining when the weekly training load is high 2 The strain is calculated in combination with Training load and Monotony and is a method for monitoring training when players are undertaking high training load and the recovery beco
158. ubmitted as the same day they are reporting As soon as the clock passes twelve 0 clock midnight the data is reported as the day after We implemented a new feature allowing the player to mark the RPE or wellness report as the day before The players can mark their response as yesterday s report before they submitting the data The reason we are allowing a player to report for yesterday 1s NIH wanted to limit the maximum exceed days to be maximum one day They meant that waiting more than a day to report data does not give accurate enough data This option is only presented when a player executes RPE or Wellness survey and not for injury survey The result is presented in Figure 4 12 The reason injury does not need this feature is because injury reporting is valid for a week and it s not dependent to report on a certain day 4 3 4 Injury reporting We are following the methodology discussed in Section 2 2 where an injury or illness reporting should be reported twice or more if he or she have multiple injuries or illness For instance if we felt pain in both knee and arms how could we report this in Ohmage We told every users in person to remember twice or more and they just have to remember it We also added an extra prompt in Ohmage where we 44 sRPE after session Summary NOTE Please make sure answers are correct before sending Choose an event Team session Choose type of session Strength training Enter duration i
159. ue Sleep Soreness Stress Mood Mood O Average 7 7 6 6 5 5 4 4 3 2 2 1 1 0 14 15 16 17 18 19 20 2 13 15 17 o Fri 17 Apr 29 03 02 04 14 04 28 04 a a Latest Wellness report b Wellnes data 30 days back Aver c Severity score three months back age Figure 4 20 Visualization of Wellness 4 3 12 Glossary A huge dilemma with the first version of pmSys was whether we should have the survey presented in English or Norwegian Ohmage did not have any support for translating questions and answers in the application In the first version of pmSys we ran two campaigns for each team one in Norwegian and one in English By creating two campaigns makes the analyzing and parsing part of the data even harder We have to fetch data for both campaign in english and norwegian and merge the data However we have removed this approach in pmSys and we only present one of the languages Instead of providing two campaigns we provide a glossary feature in the application allowing the user to check for definition of certain words see Figure 4 21 The glossary 1s hosted on a external webserver and can easily be updated as we want 4 4 Evaluation amp discussion 4 4 1 Hybrid application If we were not dependent on using native mobile functionalities like recerving push messages and the application to work offline A web application hosted on a web server would be enough and easier to develop maintain and distribute Most of our use
160. us on monitoring and analyzing The Norwegian national team and the Norwegian top division Tippeligaen used this system to monitor their players We believe this system can help sport athletes to reduce the number of overloads injuries and reduced performance in a game for the future 11 Contents 1 Introduction al Backs TOlund e y Sur Ae Lae sia ed Vay ah as tote ESS Ad ate ios oe se l2 Problem denmition ee A IA bed a N EE Cimos assii EE EN di A A EE EEE 6006 1 31000 EE uds Y do A EE EE LS Maa COnmbuuiOns sa ask folke FE Se GST ie eee hete GS tO OUM sara ess RAGE Sen SE SG Et Si d 2 Background and related work Zl System TE QUITEMENTS saa aat aS Ga SAS SEN NESA SR Ge bre 211 E ictional tequitements 1 22 5 pel al Shoe Gases kn 2 1 2 Non functional requirements LL av o 22 Sports DACKSTOUDA gs a em oh Se ess L k e e GEN Bees BESS s 2 2 1 Rating of perceived exertion ss ic sa arv vr rn nrk rn nr ren D2 VU seen Lee SPS GE es 2 2 3 Monitoring illness and injury 0 eee eee ee 2 5 Participatory SEP s Lakk Ses See FASE Fet a ee sur 24 pen mH 2 co rar A ee ee SE Ree T 2 5 District Health Information System 2 arr vr vr vr a ZO TOMES DOTE SYSICMIS 4 sa pr ke PES SE FEE SES kn G Ge Det MOSS re AI SL a EN 220 SUM 2 wes sap case Gees se EGEDE s l SE FEE See Sa bed 3 A player monitoring system using Ohmage S SVStSMAISCUSSION y dad TRER PSTN GADE Ge ON NST Bd ee ok ee A e
161. uting support we can define our own routes and make the application feel native and do not need to rely on a web server to provide us correct resources HTML template syntax we can write expression in the HTML files and then the view is rendered based on defined conditions Form validation We can for example ask a number presented in a form and then Angular can validate the numbers written in the form HTML directives We can teach AngularJS new self defined HTML directives syntax Dependency Injection One of the greatest feature with Angular it provides dependency injection across all services and controllers It exists other frameworks that are similar to AngularJS such as EmberJS and Backbone but since Ionic need AngularJS to function we did not evaluated other JavaScript frameworks and used AngularJS as the core JavaScript framework in our application 39 Supporting JavaScript frameworks X2JS When we fetch a campaign from Ohmage server the data from the server is specified in XML format With a JavaScript library named X2JS we can convert XML notation into JavaScript notation This library make the processing and working with the campaign and XML a lot easier StringJS Working with text strings is yet another challenge in JavaScript StringJS provides every missing string manipulation features in JavaScript For example native JavaScript does not support for whether a string contains in another string Underscore With
162. validation analysis and presentation of aggregate statistical data tailored but not limited to integrated health information management activities 27 See figure 2 10 a In general DHIS2 is a generic tool rather than pre configured database with an open meta data model and a flexible user interface that allows the user to design the contents of a specific information system without the need for programming background 27 Some key features of DHIS2 27 e No need to have programming experience in order to start using DHIS2 Customization and localization is done through DHIS2 s user interface In addition it can be easily set up in different countries regions or district e It can provide data entry tools which can either be in standard form predefined and tables or it can be custom made to replicate forms e DHIS2 can provide validation and improvements of data quality e Flexible and dynamic on the fly data analysis in the Data Visualizer 14 e User management module for passwords security and fine grained access control user roles e Messages can be sent to system users for feedback and notifications Messages can also be delivered to email and SMS e Integration with other software systems and use the DHIS 2 Web API and the Integration Engine DHIS2 is designed and created for clinics and hospitals where they need a system to keep personal health data organized and stored such as personal journals DHIS2 pr
163. variables into one graph and provide a scatter plot and a density plot See Figure 3 14 We did not found the charts any useful for monitoring and analyzing a player The coaches wanted graphical charts where the numbers were calculated before presenting it For example we could visualize a player s load based on the reported workload multiplied with the duration in this way we provide a better graphical analyzing feedback of the player s load and would be easier to interpret and compare various numbers for one player against another The feedback from the coaches who used the first version of pmSys said that they wanted the portal to visualize data that can be used to monitor load wellness and injury for a player rather than statistics on total responses and who have reported what We want to give the coaches better presentation of collected data 3 45 User management The user system in Ohmage was one of the main reasons we wanted to use Ohmage platform as a player monitoring system The web portal provided an easy and very powerful user management to create users and assign various classes and campaign permissions We could also change password for the users through the portal In Figure 3 15 we presented a screenshot of two classes in the portal l Comma separated values 29 Username Total Responses Private Responses Shared Responses sif Owb16j 130 130 0 sif 1396v0 127 127 0 sif 2bz8jw 121 121 0 sif 2cvdop 128 128 o sif 32ho
164. very time a campaign is altered and the language is either eng or nor We used the language definition to create the campaigns in various languages It 1s important that the team name in campaign URN is the same in the class URN Because we are using the class and campaign urn to find and validate which team a user belongs An example of an URN urn campaign demo players 6 eng A system administrator can only upload the XML defined campaign We are not allowing the coaches to do upload campaigns because we want the coaches to focus on monitoring and analyzing the player s collected data rather than administrative tasks campaign has one running state running or stopped When a campaign is stopped it is not possible for the users to submit data to the campaign However still possible for players and coaches to read reported data from the campaign If the campaign is set to running only then users can report data In additional to the running states each campaign has a privacy states which is shared or private We are not using this feature to define anything in pmSys Nevertheless we have set the privacy state to be private to ensure that shared data 1s not possible This also allows us to keep data isolated and avoid players from accessing each other s data In Figure 3 2 we have presented a small example on how a campaign is defined in Ohmage lt xml version 1 0 encoding UTF 8 gt lt campaign gt lt campaignUrn gt urn campaign de
165. vice html apple_ref doc uid TP40008194 CH100 SW9 Accessed 2015 03 01 Google cloud messaging for android Accessed 2015 03 01 Nodejs home https nodejs org Accessed 2015 04 06 Nodejs about https nodejs org about Accessed 2015 03 10 Npm home https npmjs com Accessed 2015 04 07 Modulus nodejs and expressjs sessions http blog modulus io nodejs and express sessions Accessed 2014 08 02 49 Mysql what is mysql http dev mysql com doc refman 5 5 en what is mysql html Accessed 2015 05 01 106 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 Ohmage server about the client server protocol and system entites https github com ohmage server wiki About the Client Server Protocol and System Entities Accessed 2015 02 03 Ohmage github repository https github com ohmage server Accessed 2015 04 12 Ohmage api top level entities https github com ohmage server wiki APIs for 2 x Top Level Entities Accessed 2015 05 01 Nick Antonopoulos and Lee Gillam Cloud Computing Principles Systems and Applications Springer Verlag London 2010 Peter Mell and Timothy Grance The nist definition of cloud computing Technical report Computer Security Division National Institute of Standards and Technology 2011 Oracle java platform enterprise eidition notes http www oracle com technetwork java javaee documen
166. vurdering hvordan vil du rangere M fylles ut 1 Hvilken bakgrunn har du Studieretning Yrke 2 Driver du aktivt med sport i fritiden Merk av for alt som passer Ja Nei 3 Brukervennligheten til pmSys Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 4 Brukervennligheten til Ohmage Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 5 Designet grensesnittet til pmSys Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 6 Designet grensesnittet til Ohmage Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 7 Navigasjonen i pmSys Hvordan er det man vrere i applikasjonen Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 8 Navigasjonen i Ohmage Hvordan er det man vrere i applikasjonen Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 9 Hvordan presenteres innholdet i pmSys Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 10 Hvordan presenteres innholdet i Ohmage Mark r bare n oval D rlig Ok Bra Veldig bra Utmerket 11 Hvilken applikasjon foretrekker du Merk av for alt som passer Ohmage pmSys 12 Har du kommentarer til pmSys eller Ohmage Drevet av Google Forms B 2 Rating of pmSys 91 Brukerundersokelse PmSys brukerundersgkelse M fylles ut 1 Hva synes du om prosessen for RPE wellness og injury rapportering Ma
167. ware storage or underlying infrastructure An example of this is web servers which serves HTML pages e Platform as a Service PaaS where we can host our application using the supported build environment from the provider allow the developer control the application and create the developer s own configurations to the application But the developer can not control the operating system and hardware and the underlying infrastructure e Infrastructure as a Service laaS where we are allowed to manage and control the operating system storage and applications but not control underlying infrastructure cloud structure According to Computer Business Revieew the most popular laaS is Amazon Web Services and Microsft Azure In additional to Amazon and Azure we wanted to try the cloud hosting service Digital Ocean The requirements for running Ohmage application was Tomcat servlet application and a MySQL database to store data Following java EE Enterprise Edition documentation recommended system requirements for running a Java server is I GB memory for non windows platforms 2GB for windows platforms and at least 500 MB disk storage 55 MySQL did not have any requirements but we only wanted to measure the network latency to these cloud services and therefore data storage capacity it s not considered Since we want to measure network latency to various cloud services and compare to Troms s server we should pick the servers that are cl
168. when we need to host the data ourselves and be the owner of the data or 1f we want to have low latency by having the server close to our location We have also presented four new improvements for the server and this was done to prove that it exist many improvements for the server 80 result success data urn ohmage context timestamp values 2015 04 25 21554328 4 2015 03 31 18 26 20 2015 03 24 22 27 52 br urn ohmage prompt id srpeLength values 90 90 93 ly context display_label Duration id srpeLength prompt_type number default lt 90 ss and more F urn ohmage prompt id srpeWorkload values 0 1 4 ly context display_label Workload id srpeWorkload prompt_type single_choice ics and more F urn ohmage prompt id srpeType2 values NOT DISPLAYED NOT DISPLAYED NOT DISPLAYED l context 4 display_label Session condition srpeTypel 1 or srpeTypel y id srpeType2 prompt_type single_choice and more Fz urn ohmage prompt id srpeTypel values 0 Oy 0 ly Context 4 display label Session id srpeTypel prompt type single choice sss and more Figure 6 7 Example of response read in Ohmage result suc
169. xt plain res end Hello World n listen 1337 127 0 0 1 console log Server running at http 127 0 0 1 1337 3 Figure 5 1 HelloWorld code in NodeJS listening for HTTP connections The idea is to have NodeJS as a server instance running and fetching data from Ohmage server over HTTP protocol and serve like a proxy instance between the client and the server In NodeJS we can also have HTML server side rendering to present content This approach is different from normal single page applications where the HTML rendering and HTTP fetching is managed by the client A system model of our envisioned web portal is presented in Figure 5 2 Ohmage server Web portal q Tt D o N gt Database Web service Node Client Web portal Web browser Figure 5 2 pmSys web portal System overview 5 2 2 Web application According to NPM s homepage it exists about 138487 packages libraries and frameworks 47 to NodeJS Like every other programming languages projects every project have a project file to resolve I ACID Atomic Consistent Isolated Durable 59 dependencies NodeJS also have a project file structured in a JSON format To have a project file where we define all dependencies our project need in NodeJS is one of the benefits of using NodeJS ExpressJS is a web framework to NodeJS and serves as a layer on top of NodeJS ExpressJS ease the web application development and let the developer wri
170. yers where unable upload responses to the server and the coaches could not fetch data from the server Hence we started to consider cloud hosting since the cloud infrastructure upgrade is a less problem at the cloud services and they can guarantee up to 99 up time 6 2 Ohmage server In this section we are going to look at the server and see how the server 1s built Ohmage back end consist of various applications and built like stack with application layers All information about the users teams user roles campaigns and responses are stored in the database layer of Ohmage back end The database layer in Ohmage is a MySQL database MySQL 49 is a Relational Database System which means the data is stored in rows and tables Ohmage also supports MariaDB which is a improved version of MySQL Ohmage database contains many different tables and we will not present them all here but some of the most important tables we are using is presented and described in Table 6 1 I MySQL vs MariaDB https mariadb com kb en mariadb mariadb vs mysql features 71 72 Table Description preference Meta information about the back end instance user User information such as username and password class Information about a class team Table 6 1 Few tables from Ohmage database Above the database layer we have an application layer The application layer is built to prevent users from accessing the database directly Allowing a user to insert dir
171. yers with pen and paper is nearly impossible to organize and analyze reported data Alternatively consider web survey services such as Google Forms to collect data through questionnaires Regardless of these two methods the data must be manually processed and analyzed before presented and evaluated With use of modern technologies monitoring becomes more efficient and time saving We can for example have a mobile phone to collect data and then automatically send the data to a distributed server and the server automatically processes the captured data and stores 1t securely A web application can fetch the data from the distributed server graphically presenting and visualizing the captured data with charts This envisioned system does not only simplifies the process of monitoring sports athletes but saves time and let users focus on analyzing and evaluating collected data It is also possible extend the system to interpret and analyze captured data by itself For example implement machine learning and let the system tell when a player 1s vulnerable to injuries or have reduced performance in the next games In this thesis we investigate and research an implementation of such monitoring system by self report of subjective data We will work in close cooperation with PhD student H vard Wiig and Dr Thor Einar Andersen from NIH to see if our envisioned system can support their research project I Google Forms https www google com form
172. ystem by lowering the network latency and adding new endpoints to reduce data transfers between the server and client To summarize we have implemented a working system that is used in real life to monitor football players We have implemented an effective way to collect data simplified and removed the process of parsing and processing captured data manually We have also created presentation tools to visualize and present captured data where users can analyze and evaluate captured data 1 6 Outline We begin this thesis with presenting requirements and related background in chapter 2 such as sport related backgrounds and other related systems and concepts In chapter 3 we investigated and studied Ohmage as a platform for pmSys Ohmage is a data collection platform through self report We then Ny app skal sikre EM suksess NRK http www nrk no troms ny app skal sikre em suksess 1 12048486 accessed 2015 05 03 evaluated the existing mobile application and web portal as monitoring tools We concluded the chapter with an explanation on why Ohmage did not fulfilled our purpose In the next three chapters we have presented three different applications used in combination to form a player monitoring system In chapter 4 we present pmSys mobile application created for athletes in sport context Here we present technologies we used to implement the application Furthermore we evaluated the new implemented application against the existing applic
173. ystem hosted in various cloud services 32 Chapter 4 The pmSys mobile application In this chapter we present the new mobile application we implemented as a replacement for Ohmage mobile application We are going to present technologies we used to implement this new application and the improvements we talked about in the previous chapter We also present two user studies we performed to confirm whether pmSys mobile application was better or not Then we present the result of the new implemented application 4 1 Motivation In the first version of pmSys we distributed the system with Ohmage mobile applications as a self reporting tool With the improvements of Ohmage applications from the previous chapter a new mobile application was highly demanded We wanted to improve the reporting process for our users We also wanted to motivate the players to use the application by visualize captured data with graphical charts instead of showing raw response as in Ohmage The players tend to forget to conduct their survey especially after training and in the mornings Providing a possibility to add reminders helped to a point but not everyone was able to or knew how to add reminders What 1f we could have the coach to send instant message to their players to remind each player of conducting their survey What 1f we could bring this idea further by letting the coach create a scheduled reminder that can automatically send a message to all the player

PmSys: a monitoring system for sports athlete load, wellness & injury

Contents

Download Pdf Manuals

Related Search

Related Contents

PmSys: a monitoring system for sports athlete load, wellness &amp; injury

Contents

Download Pdf Manuals

Related Search

Related Contents

PmSys: a monitoring system for sports athlete load, wellness & injury