Home

Health Monitoring Analytics – Final Report

1. It links to the DisplayChart class to display different types of charts for users Controller DBConnection Checkidentity CreatePersonalProfile UpdatePersonalProfile AccessDisplay MessageManager AceessDisplay is ValidLocation STATE_NON_LOGINiint 0 STATE_LOGGED_INiint 1 STATE_REGISTERING int 2 STATE_RETRIEVE_ACCOUNT int 3 STATE__ADMIN_MAin int 4 STATE_VIEW_VIEW_GMAPiint 5 STATE_VIEW_CHART int 6 STATE_SEARCH_HOTCITY int 7 STATE_SEARCH_FACILITY int 8 STATE_SEARCH_EXERCISE int 9 STATE_DATA int 10 STATE_Chart_line int 11 STATE_Chart_Fan int 12 STATE_VIEW_Cloud int 13 Figure 24 Logic Diagram of the VGA Controller Block 10 3 Traceability Matrix Domain Concept UC1 UC2 UC3 UC4 UC5 Controller x x x x x Interface x x x x x Data Analysis x x x x Communicator x x x x x DB Connection x x x x x Table 10 Traceability Matrix 11 System Architecture and System Design 11 1 Architectural Styles In our software there are three abstraction layers being implemented in our design e User Interface 68 This is top level of the software system and is the only part the user is aware of It displays all the relevant information that the user needs when searching the site inputting queries et cetera This tier communicates with other tiers by sending results to the browser and the other branches of the network A website for
2. Failure The website could no display the word cloud or only some of them The word cloud is not dynamic which could not move with cursor Failed to connect with the twitter or connect to irrelative info e UC 15 Viewing Line Chart 88 Success The website display line chart integrated and clearly The percentage of different US regions could be displaced as tendency with different time to illustrate the change in one day Failure The website could no display the line chart or only some of them The line chart is not dynamic which could not show the tendency of one day e UC 16 Viewing fan chart Success The website display fan chart integrated and clearly Failure The website could no display the fan chart or only some of them 15 History of Work Current Status and Future Work 15 1 Plan of Work The Plan of Work is illustrated in the two figures below Task Name y Duration Start y Finish JE 2013 f ctober 2013 November 2013 10 13 16 19 22 25 28 1 4 7 10 13 16 19 22 25 28 91 3 6 9 12 1 Proposal 6 days Mon 10 14 13 Mon 10 21 13 a Y First report 18 days Sun 9 22 13 Tue 10 15 13 Y part 1 9 days Sun 9 22 13 Wed 10 2 13 O gt _ Y Caine Statement of Requirements 4 days Sun 9 22 13 Wed 9 25 13 CSR Y Systen Requirements 4 days Thu 9 26 13 Tue 10 1 13 E Y Project Management and References 1 day Wed 10 2 13 Wed 10 2 13 a Y Part 2 5 days Thu 10 3 13 Wed 10 9 13
3. Twitter twitter twitter etc So even if we search for twitter we re still going to introduce noise that isnt related to our hashtag We can search for multiple hashtags with this it s the primary parameter for our search e Locations This is a CSV list as well of parameters The Twitter API parses in sets of twos to look at the latitudes and longitudes The API takes four GPS coordinates two latitudes two longitudes and uses those to create a bounding box as shown below 46 The bounding box has the two GPS pairs as red dots in the corners We capture the black point inside the red filled box and ignore everything outside of it Bounding box is NOT used in conjunction with other filters Using location and track together say location is set to New York and track is looking at twitter then we would get all tweets related to New York and all tweets related to twit ter So you ll pull about 1 000 tweets from NY about things totally unrelated to twitter and then youll pull all tweets worldwide as well that only deal with twitter e Delimited This parameter lets you set how long you want the tweet to be that you re search ing for e Stall warnings This is a logical yes or no that just sets whether or not you want stall warning set This sends messages back to you about whether or not you re about to be disconnected People using the streaming API have their bandwidth limited
4. day Int Password String LocalDatabase Figure 23 Logic Diagram of the VGA Controller Block 10 2 5 LocalUser LocalUser is a derived class that represents a users information inside of the database Most of the private attributes in this class are personal information that has been described in previous sections e Friendlist A String pointer that points to the beginning of a linked list of userIDs the users friends e Password A String variable that is a data member of class LocalUser Password variable as seen earlier e LocalDatabase This is an operation that grants access to the system when it makes a call to open a portal between the database and the website 10 2 6 TwitterData TwitterData is a derived class used when requesting access to the Twitter Database The functions of this class are below 64 e AccessTwitterDB This operation is used whenever we need to access the Twitter Database and obtain data from it Obtained data will be used in other classes and operations 10 2 7 Search This class is meant to implement the functionality behind the searching function in the application The data stored in this class is e hotCities This is a String variable that contains a list of all the hot cities requested in a query healthSuggestion This is an as of yet unimplemented class that would contain a series of sugges tions on how to improve the users health
5. 10 2 2 DisplayChart The DisplayChart class implements the main functionality for the chart and map user inter face elements It contains the following variables e googleMapChart This is a MapFragment variable that is stored directly inside the class MapFrag ment is coming from the Google API e Color This is a String variable that holds the value of the color the user wants the map markers to display The function of this class is listed below e display HotCities googleMapChart color The user will input where they want to look and what color they want the markers on their map to be A MapFragment is then dynamically generated updated to match their preferences and passed to this displayHotCities function This will propagate through the system to the output 10 2 3 AccessProfile When dealing with the use case where the user needs to modify their personal information an additional class needed to be introduced to handle the information outside of what is needed for simple registration e g name and email As such we have this class AccessProfile These include various set and get functions for class encapsulation AccessProfile exists as a class since it needs to be instantiated for more than one user who accesses their profiles at a time The variables for the class are listed below e Username String variable corresponding to the users username e Password 99 String variable correspond
6. Boolean STATE_REG NGhnt password String setAdmin admin boolean void STATE_RETRIEVE_ACCOUNT int 3 age int equals objectObject Boolean STATE__ADMIN_MAin int 4 sex Char addUser SnewUser void STATE_VIEW_VIEW_GMAP int 5 year int deleteUser Susername void is ValidLocation month int STATE_VIEW_CHART int 6 day int STATE_SEARCH_HOTCITY int 7 email String DisplaC hart STATE_SEARCH_FACILITY int 8 setAge int void setAge int STATE_SEARCH_EXERCISE int 9 createMap setSex Char void STATE_DATA int 10 ees enterMapTA DEE STATE_Chart_line int 11 setNewName String void clearMap STATE_Chart_Fan int 12 setNewPin String void 9 gleMap STATE_VIEW_Cloud int 13 RETA ieChart gt gerNewPin String pie a setYear int void sport String getYear int Data hotCities String VER getMonth iint Line chart healthSugguestion setDay int void displayHotCities tweets String getDay int exercise String setEmail String googleMapChart color facility String getEmaill String displaySport sport void setHotCites String Displaydata data void setFacility facility String void manageinbox void getFacility String chageProfilePicture void Displaylinechart linechar void i i getLocation String DisplayCloud data void getExercise String DisplayFan data void getTweet string Figure 22 Class Diagram 56 The figure illustrates most of classes
7. HEALTH MONITORING ANALYTICS FINAL REPORT SOFTWARE ENGINEERING GROUP 1 EDITED BY GRADEIGH D CLARK XIANYI GAO RUI XU LI XU YIHAN QIAN XIAOYU YU Rutgers University School of Engineering DECEMBER 20TH 2013 HTTP WWW HEALTHMONITORINGANALYTICS COM HTMLApp2 PUBLIC_HTML INDEX PHP 1 1 Individual Contribution Breakdown Task Xianyi Gradeigh Rui Li Yihan Xiaoyu Total Summary Of Changes 5 16 7 16 7 16 7 16 7 16 7 16 7 100 points Sec 1 Customer State 16 7 16 7 16 7 16 7 16 7 16 7 100 ment of Requirements 6 points Sec 2 Glossary of Terms 4 16 7 16 7 16 7 16 7 16 7 16 7 100 points Sec 3 System Require 16 7 16 7 16 7 16 7 16 7 16 7 100 ments 6 points Sec 4 Functional Require 16 7 16 7 16 7 16 7 16 7 16 7 100 ments Specification 30 points Sec 5 Effort Estimation 4 16 7 16 7 16 7 16 7 16 7 16 7 100 points Sec 6 Domain Analysis 25 16 7 16 7 16 7 16 7 16 7 16 7 100 points Sec 7a Interaction Dia 16 7 16 7 16 7 16 7 16 7 16 7 100 grams 30 points Sec 7b Design Patterns 16 7 16 7 16 7 16 7 16 7 16 7 100 10 points Sec 8a Class Diagram and 16 7 16 7 16 7 16 7 16 7 16 7 100 Interface Specification 10 points Sec 8b OCL Contract 16 7 16 7 16 7 16 7 16 7 16 7 100 Specification 10 points Sec
8. e storeNewPin pass This function would be called internally within any setNewPin function and would update an associated value e g password within in the DB for that update e storeAge Int This function would be called internally within any setAge function and would update an associated value e g age within in the DB for that update e storeSex char This function would be called internally within any setSex function and would update an associated value e g sex within in the DB for that update e storeDOB year month day This function would be called internally within any call of set Month setDay set Year function and would update an associated value e g day month year within in the DB for that update e storeEmail email This function would be called internally within any setEmail function and would update an associated value e g email within in the DB for that update As shown in Figure 3 2 class Database has two derived classes LocalUser and Twitter Data These two derived classes contain lists of data for other classes 63 storeNewName String void storeNewPin String void storeAge Int void storeSex Bollean void storeDOB Int Int Int void storeEmail String void getLocation String LocalUser TwitterData TwitterData useriD String AccessTwitterDB username String email String sex Boolean age Int year Int month Int
9. FFworkout strengthtraining FFsmashfit FFsquash swimming biking Fey cling losing weight Akeepfit Table 9 Table of Used Hashtags These tweets were selected by looking at the hashtags correlated with tweets about fitness Since we are not limited by the number of hashtags we grabbed any that seemed relevant Problematic hashtags are ones that seemed good but ended up generating junk data health always seemed to correlate to news results or current events e g bioterrorism Never saw any tweets with health that had anything to do with our study train was also similar it generated results about trains and accidents that occurred recently or people complaining about being late for the train So this is an issue with context As for when we do the search we get a list of hashtags associated with the tweet every hashtag the user used So if we don t know what hashtags to look for we only need a few health related ones e g fitness and then we would get tweets from users that are relevant with fitness and would contain many more hashtags with it that were likely relevant because the users tend to group their tags in this way heuristically So one can increase the sample space of where to search for by increasing their tag number of course at some point increasing hashtags in this way must result in no returns at all We imagine the curve would be something like a decaying exponential that rise
10. https dev twitter com 5 Engrammi http engrammi blogspot com 2011 10 nanni balestrini original tag cloud html 6 Where DO Twitter F Bombs Come From Heat Map Shows Rudest Places http mashable com 2012 08 22 twitter rudeness heat map 7 Healthcare Social Media Analytics http www symplur com healthcare social media analytics 8 HMS Health Monitoring Systems http www hmsinc com media EpiCenter_2 17_User_Manual pdf 9 Sylwia Krol Bozena Zabiegala Jacek Namiesnik Monitoring and analytics ofsemivolatile organic compounds SVOCs in indoor air Anal Bioanal Chem 2011 400 1751 1769 DOI 10 1007 s00216 011 4910 x 10 Monitoring the health of web page analytics code http www google com patents US20110035486 91 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Apoorv Agarwal Boyi Xie Ilia Vovsha Owen Rambow Rebecca Passonneau Sentiment Analysis of Twitter Data Department of Computer Science Columbia University New York NY 10027 USA fapoorv cs xieQcs iv2121 ram bow ccls becky csg columbia edu Twitter Data Analysis Gaurish Chaudhari Under the Guidance of Prof Sunita Sarawagi http www cse iitb ac in gaurish Seminar_Presentation pdf Kenneth M Anderson and Aaron Schram Design and Implementation of a Data An alytics Infrastructure in Support of Crisis Informatics Research Proceedings of the the 33rd International Conferenc
11. 10 260 Twitter Data si Sede e a Jade Ok ay rate a ela ie d en ean lias OB ni Arii 20 TDI Barc ais NA hed A ADD De oe ue a ls 10 28 See EE ITA CORTO lle gt bad a clog Sale feo ei AE A ale ets bee fe ne 10 3 Traceability Matrix A d fale hot ge at BG ond Boe ale eles 11 System Architecture and System Design DMA geheelt e Gy 0 Rng eek SR gi A A A 11 1 1 Browser Server Structure and 2 Tier Architecture 11 12 Tier o a ere eae et gan a de Rete 11 2 Identifying Subsystems daa io deren aoe ade EE 11 3 Mapping Subsystems to Hardware 22 22 0004 Lica Network ee dato gate et oe dhs oe ee BAP EE poke EEN 11 5 Global Control Flow se 4 04 a e e as a 11 6 Hardware Requirements 46 2 6 sy ct oa ode Dhaene eke Behe Ce hos Sy oe ate ee S 12 Algorithms and Data Structures lr A node We a We SB ek wt RO O 12 1 1 Estimation of population concerning health 12 1 2 Word Cloud Alora die Bina OR A a ale eS ai 12 1 3 Database Querying for Word Cloud 12 2 Data Str ctire sate g e chee aa a el ee 13 User Interface Design and Implementation 14 Design of Tests AT AS ea i Se ed ee es Se ee Sek eee es 14 1 1 Deprecated Use Cases o e e eee eee a o AA e a A BL aed d AALS ae ath ts 14 2 1 Deprecated Cases y o oot he Gee al Moe ie Pal eS BAER iai 15 History of Work Current Status and Future Work Ee Worker a aaa da ia Renae RS aa eh nae ee 15 2 Project Coordination and P
12. 3 Viewing Results Entering Webpage Process Data Requesting Tweet data Requesting Tweet dath Total Tweets Counting US tweets counti and calculating fo estimated number pe opl H Controller AccessDisplay Request for dsplay Process Notification Button Display Results in Notification Butto 49 Figure 18 Use Case 3 View Results Database getTotalTweet Use Case 3 is designed with the intention of giving the user the ability to view the results the system has to offer on its data processing Results here refers to displaying how many tweets have been aggregated by the system how many tweets have a proper geolocation how many tweets are from the United States and finally how many tweets are in each state The interaction diagram displays the process the system needs to go through in order to handle the user s input and correctly display the output it waits for a user request and then generates the correct chart type off of that The user needs to navigate to their index page to see the analysis in a more direct fashion The user will see pie charts that split the tweets into four regions of the United States and those four regions are further broken down into the later 50 states that make them up For the result displaying we also added a statistical analysis graph counting the number of tweets produced over 24 hours of day The tweets we obtained also include the time when a tweet is posted This giv
13. Cloud function 1 NAVIGATION total 5 mouse movements as follows e Move cursor to the chosen word cloud e Move cursor to find the chosen word if not in this page Scroll mouse Click the chosen word cloud 79 Click interested info N Function display Move mouse to see the word cloud move with the cursor Scroll mouse the word cloud would change its size by shrinking or magnifying Provide extra explanation with each word cloud to connect with Twitter Markers aw ER Map Satellite Loge WW Map data 2013 Google INEGI 500km i TemsofUse Figure 33 Markers Markers w CN mme ane ax A New Holland Chester lensid INB N Bristol Flo ip __ p Map Satellite TO fe a LAA E Si we a SCH ilad iph or on N py 3 e wnship Gym bagi Getting hypedddddd gym goingtothegym goingtothegymtoday http t co Wwt39x21kV Pennsylvania New Hanover Township a EE S Downing E Coatesvill 4 D I parkesburg La Wolf s Hollow county Park Cochranville N K i 3 A d d D t A poli de A La S ay b Sal le y Grove i Bra dywine Boot y L 4 race im d US 4 ite A WOM Sar AAA 5kmi TemsofUse Reportamapemor n Figure 34 Markers Example Running Viewing map function 1 NAVIGATION total 3 mouse movements as follows e Move cursor to the chosen place Scroll mouse Move cursor to chosen location N Function display Move mouse to see the
14. No precondition POST CONDITION The user could achieve results they want Flow of events 1 The user visitor or member types in his interested location facility 2 The system requests for the data from Twitter database about the tweets from a location or about some facilities information 3 The raw data get analyzed in data analysis component of the system 4 The statistical and graphical results are shown in the interface page e UC 2 Viewing Statistical Breakdown Initiating Actor Member 21 Actor Goal To view statistical breakdown Requirements REQla REQ1b REQ3 REQ6 REQ7 PRE CONDITION Data must be present in the database POST CONDITION The information is displayed to the user Flow of events 1 User goes to the main page to view the data 2 Data if present in the database is pipelined to the user 3 Charts are populated and presented as a breakdown UC 3 Viewing Results Initiating Actor Member Actor Goal To view results Requirement Addressed REQ1b REQ2b REQ3 PRE CONDITION Data must be present in the database POST CONDITION The information is displayed to the user 1 User goes to the main page to view the data 2 Data if present in the database is pipelined to the user after filtering 3 Results buttons at top are populated and presented numerically UC 4 Viewing Map Initiating Actor Member Actor Goal To view maps Requirement Addressed REQ1b REQ1c REQ2b REQ3 REQ4 REQ5 PRE CONDITION Data m
15. Since we are using firehose on the streaming API we grab as many tweets as we can before Twitter truncates our connection Ideally we would keep it open forever but at some point since we are not business customers we cannot make heavy demands on their servers for very long Connections can be unlimited but the bandwidth used by it is capped a status warning would send typical warning back to you about how much percent you are towards getting cut off so you can adjust your stream requests on the server end Since we are running a Linux server on our host we have access to what is called a cron job A cron job is just a script that is scheduled by the server to occur at periodic intervals by invoking the cron command on the PHP script So the task can be scheduled by our web host to run two times an hour per any job at a specified time xx 05 amp xx 15 for example so twice an hour at 5 minutes after and 15 minutes after However we can schedule multiple cron jobs to occur per hour So this means the tweet retrieval script can be run once a minute if we schedule 30 jobs that occur at xx 00 xx 01 and xx 02 xx 03 et cetera Our setup is two cron jobs to do the tweet retrieval that run at xx 00 xx 15 and xx 30 xx 45 More can be scheduled but it didnt seem worth the benefit because we can retrieve tweets we already have before it is a delicate balance 47 The alternative was to keep the stream open as long as possible by having the ser
16. This is all well and good But we know that many of you have voiced that you lack the social structure that would encourage this type of domino effect More specifically you may lack friends or acquaintances who exercise daily and eat right or have not been impressed upon by the friends you have about the importance of exercise in a fulfilling life So the question is no longer How do I exercise but How can I meet people who exercise and How can I find out how much of my community is exercising This additional question is important because we are at our cores competitive creatures You are willing to better yourself to match your community if you can And this finally is where we step in to help you the customer To solve your problem we plan to design and implement software to track the health activities of your community analyze the progress as a whole and feedback the analysis to you as a motivator We intend on keeping track of whether your community becomes health ier or at least more physically active as a whole by aggregating any and all available data in your neighborhood and surrounding communities And there already exists a veritable ocean of information to draw from Twitter This real time social media engine generates tweets by the thousands per minute on any variety of topics the goal would be to focus on the health related aspects and localization of the tweets With the big data analysis we would be hoping to target some o
17. User Customer A person who uses the Health Monitoring Analytics software Administrator The person who is in charge of maintaining the software Tweets 140 character statements delineated by hashtags that we are using to analyze the health of a community Database This is where the tweet information is stored after being pulled from Twitter Real time Graphics Used to illustrate aspects of regions of the United States as a function of how the tweets are distributed Analysis types include tweet aggregation by hour of day location or distribution Tag Cloud A pictorial representation of the most frequent hashtags using the size or color of the words as a weighting factor Heatmap A pictorial representation of the distribution of the tweets by grouping close by tweets together and assigning warmer colors to highly grouped messages 13 5 System Requirement Analysis 5 1 Functional Requirements Table ID Priority Weight Requirement REQ 1a 5 System should retrieve data from Twitter REQ 1b 5 System should retrieve data from database REQ 1c 5 System should retrieve data from Google REQ 2a 5 System should filter data by hashtag REQ 2b 5 System should filter data by location REQ 3 5 System should store Twitter data REQ 4 4 System should display the distribution of data via a heatmap REQ 5 4 System should display all relevant tweets as markers on a map REQ 6 4 The s
18. emcee SY de eens Se Sek 18 05 Use Caserta de eee aS aie ue AA ee Behe e IN Bok od ree Bi 19 6 3 1 Casual Descriptions vos dales ia E Rea ateu eh nd ah sens Gea t 19 6 3 2 Use Case Diagrams asi a ee Sie e eee 20 6 3 3 Fully Dressed Descriptions eect it E o 21 6 3 4 Deprecated Use Cases a 23 OA Traceability Matix s se me ie a en ia Bd Se wd de ee ew d 24 6 5 System Sequence Diagrams dl rs Goh eden ee 3 d 24 7 User Interface Specification 31 7 1 Preliminary Design ss AAA E SA ee ed 31 7 2 User Effort Estimation ooa aa eo eee 32 8 User Interface Analysis 35 Sl Domain Modelo ra a A e Soe a Be EE 35 8 1 1 Concept Definitions pta ea ee Pas a Pe 39 8 1 2 Association Definitions AAA 0 Abe ed web 35 8 1 3 Attribute Definitions son amp A Be othe BS Gass BS Gad 36 8 1 4 Traceability Mati be Pa ed a 36 8 2 System Operation Contracts had e a SS ee Se A pi 36 8 3 Mathematical Model 37 8 3 1 Before Gathering Data les E tg te Eee ru Oh Gee eH aad 37 8 3 2 After Gathering Data iba eal he Gra eee 4 Oates ta 37 9 Interaction Diagrams 39 10 Class Diagram and Interface Specification 10 1 Class EEN ee ee ER ee 10 2 Data Types and Operation Signature GER EC E EENEG lt E li a A pa e ii a ode Tale d de Dali i para E aa atata Soia 10 2 3 AccessProfile ooa A e i i ci dale Ti ee ee Te oe i pos de a 10 24 Databases Bi Beets Bete amare E aoe ote E E eG aja e ye aa a ROZ ME pa etek a E E at
19. exercise by seeing people you know working out and subsequently joining them or by meeting more people in your community who exercise as well transforming a previously solo venture into a social activity We posit the following to you there exists a definitive culture around those who exercise regularly it is not uncommon for people to be swept up in the exercise habits of their friends If your friend starts cycling as part of their commute and tells you that it gives him her clarity of mind freedom of transport and exercise as if that is an afterthought then you are more likely to also start cycling in some way There is no guarantee but it can be asserted that at least the probability that you will take up cycling as a means to better yourself Fundamentally this operating principle of exposure and awareness can serve as a hook to a healthier lifestyle Visual impacts cannot be ignored either what if strangers witness the newborn cyclist on his commute It wouldnt be wrong to assume that this could serve as the spark for an engine of change in yet another person to begin cycling If them why not me Now there is a domino effect of changing peoples lifestyles for the better What we assert now formally is the power of the acquaintance the friend to influence both you and their community at large by extrapolating the bicycle example to all types of exercise a By pO or E f f kee Figure 1 The community influencing a user
20. has to go through a loop to determine which state it belongs to See Figure 1 for an example of the bounding box for Arkansas This is plotted with one of our bounding box data Four makers indicate four corners of the bounding box We can see from the figure that one of the marker lands on Mississippi 44 CV GIS o Arkansas Z S pringfield APS 2 Y a CH Y Clark d ers i 1 RY iken Fayetteville Jonesboro j LA row A a T la o d Jackson AR Cordova Arkansas we T Little Rock y Z W Pine Bluff y g V B A Xe El Dorado Greenvill o Mi SS BA Tuscalc E a o A E dl Fi Figure 17 Example of a bounding box We couldn t design the filter to have a set of hashtags and a specific location during our data mining In other words we can choose to pull all the tweets with a set of hashtags but we have no control on where these tweets should be from Alternatively we can choose to pull all the tweets on a specific location for example New Jersey but we cannot set up the hashtag requirement at the same time Knowing that twitter API doesnt allow filtering of tweets with more than one specification we decided to pull data with hashtags so that we only extract useful data from twitter site This leaves us having no control in tweets location Therefore we got the tweets from other countries as well Using the bounding box illustrated above we extract out all the tweets from US As for th
21. have their location hidden As for explicit filtering we obtained many tweets However in our SQL table we have two entries that we can use to filter the data geo_lat and geo_long when we get the JSON file of a tweet from Twitter it contains latitude and longitude coordinates We sort the data into our tables like that with this name We can construct an easy SQL query to request Tweet data such that we are only pulling Tweets from our database into our web page that has latitude and longitude coordinates like this type of query e Slatquery SELECT geo lat FROM tweets WHERE geo_lat 0 ORDER BY geo long e Slongquery SELECT geo long FROM tweets WHERE geo_long 0 ORDER BY geo long e textquery SELECT tweet_text FROM tweets WHERE geo lat 0 ORDER BY geo long Then just execute the query via PHP commands On our server end we implement the tweet filtering via bounding box through JavaScript as explained in another answer 93 R UC 5 Viewing Word Cloud Entering Webpage Display Word Cloud Word Cloud Library Initiatiating Page Requesting Twitter EES Requesting Twitter Da e 5 Database getTotalTweet Total Tags Extraction Tag Frequency Calculation Word Cloud Plug In Display Word lt Cloud nteraci with Figure 21 Use Case 5 View Word Cloud Use Case 5 is designed to visualize the popularity of each hashtag based on the frequency of appearance in tweets The frequency
22. he may log in to the website ii Personal page This button brings the user to his personal profile page iii Account settings This button brings the user to the account settings page iv Log out This button is visible only if the user is logged in and will log the user out 15 e Google map window This is a modified map ported over from Google Maps that displays the locations of facilities and tweeters in the system f Data analysis windows This displays real time analytics of the communitys twit ter data A user can view a histogram chart etc any type of analytic data for the community g Help link This is a link that takes the user to the help screen h About link This is a link that takes the user to the about screen Log in screen This is a dialog box where the user enters his or her login name and password to gain access to his profile on the website There exists a button on this box to register for the service and this button navigates the user to the registration screen Registration screen This page is where the user enters their personal information to create an account for the software Help page This is a page where tutorials are supplied for the user to help themselves learn how to use the system About page This describes the motivation for the project Personal page This is a page where the users personal information is displayed back to them along with other such relevant
23. is visible error handling in the diagram for the instance where the strings are not valid or if they are partially complete 42 A U 2 Viewing Statistical Breakdown Entering Webpage Process Data Requesting Tweet data Requesting Tweet dath Database getTotalTweet lt eS e E a am au controller VIEW_MAP ZAS Cem e mm em Prepare Map Display Map i SE ebe oe Process Data Output results Display Pie Charts Figure 16 Use Case 2 View Statistical Breakdown 43 Use Case 2 is designed with the intention of giving the user the ability to view statisitcal breakdown data Graphical data here refers to the breakdown of tweets by contiguous regions of the United States The interaction diagram displays the process the system needs to go through in order to handle the users input and correctly display the output it waits for a user request and then generates the correct chart type off of that The user needs to navigate to their index page to see the analysis in a more direct fashion The user will see pie charts that split the tweets into four regions of the United States and those four regions are further broken down into the later 50 states that make them up Following is the description about how we collected tweet data and categorize them based on the location All of our tweets are pulled corresponding to a set of hashtags Most of the tweets that we collected don t have location data so the latitude and longitu
24. map displaying sports map 80 e Figure 11 9 Markers e Scroll mouse the map would change its size as shrink or magnify e Move sursor to the mark our website will display more info automatically Figure 37 Search Example 3 81 Heat Map vZ Map Satellite U 4 Guy i Monterte Map data 2013 Google INEGI 500kmL___4 TemsofUse Figure 38 Search Example 4 Running Search function 1 NAVIGATION total 5 mouse movements as follows i Search without key words e Move cursor to search box e Print searched place e Move cursor to map diagram ii Search with key words e Move cursor to search box e Print searched place with key word e Move back to map diagram Function display e Move cursor to the search box to input a place e View the searched map on the map diagram e Move cursor to the search box to input a place with key word e View the searched map with marked key word situation on the map diagram 14 Design of Tests 14 1 Test Cases e UC 1 Searching for Locations and Facilities This use case should enable the user to search for locations and facilities For location search the functional system should display results in the form of graphical analysis 82 visualization through a Google MapFragment with tweets labeled as red dots and statistical results with percentages showing the health population of the location For facility search the
25. map should display available facility locations across the United States showing the distribution of it over several cities e UC 2 Viewing Statistical Breakdown The user needs to navigate to the website and see whether or not the pie charts are being split into statistical values for the contiguous United States If there is no breakdown then the server is either unable to connect with Twitter Google or the website database doesn t have tweet data e UC 3 Viewing Results The user needs to navigate to the website and see whether or not the results counter is changing across an hour If there is no breakdown then the server is either unable to connect with Twitter Google or the website database doesn t have tweet data e UC 4 Viewing Map The user needs to navigate to the website and see whether or not the maps are dis playing correctly with heat on the United States along with markers If there is no breakdown then the server is either unable to connect with Twitter Google or the website database doesn t have tweet data e UC 5 Viewing Word Cloud The user needs to navigate to the website and see whether or not the word cloud is being displayed correctly If there is no word cloud then the server is either unable to connect with Twitter or the website database doesn t have tweet data 14 1 1 Deprecated Use Cases These use cases have been deprecated due after updating our goals but we leave them here on the instruction that th
26. of twitter users who tweets about health and exercise instead of the number of tweets 12 1 2 Word Cloud Algorithm As for how to algorithimically generate the size for the word cloud we heuristically assigned the following formula Size logio Frequency 1 1 x 2 4 We add a one to the frequency before taking the logarithm in the chance that there are no instances of a hashtag a 1 is added after to make the weight when multipled by 2 a little more meaningful e g the size of the text is always greater than unity at the very least so it is legible 12 1 3 Database Querying for Word Cloud The plugin we used is called TagCanvas http www goat1000 com tagcanvas php This is an HTML5 plugin that allows us to plot the hashtags onto the map through HTML The problem is that we need to generate the weights of the words ourselves We can obtain rankings of our hashtags by first querying our database like so tags array run running gym jog jogging jogger marsic exercise exercising exercises fitness fit fitfam seenonmyrun fitnessfriday fitforlife mefirst gymtime healthy 12 bodybuilding chest workout getfit active healthychoices diet cardio strong training fitnessaddict fitspo instahealth trailrunning trailrunner trailrun fitspiration fit
27. page or any page of the website where the search functionality is enabled From there they can type their query into the search bar and wait as results are compiled for them from the system There is visible error handling in the diagram for the instance where the strings are not valid or if they are partially complete 40 R UC 1b Searching for Facility Input Facility search getFacility string Controller DBConnecioni Facility data fom Googe senice combined with Facility Data fom Googe Senice No data available for this facility Process Data l Controller AccessDisplay Request for display DisplayChart createMap Display results map and chart Figure 15 Use Case 1b Searching for Facilities 41 Use Case 1b is designed with the intention of giving the user the privilege of searching for facility data through the user interface This is important for less ancillary goals of the application the ability to search for facility data is used to suggest places for the user to go rather than showing data The interaction diagram displays the process the system needs to go through in order to handle a query from the user and correctly display the output The user needs to navigate to the index page or any page of the website where the search functionality is enabled From there they can type their query into the search bar and wait as results are compiled for them from the system There
28. qy Y Functional Requirenents 3 days Thu 10 3 13 Mon 10 7 13 E Specification Y User Interface Specification 4 days Fri 10 4 13 Wed 10 9 13 E w Part 3 5 days Wed 10 9 13 Tue 10 15 13 Y Domain Analysis 4 days Wed 10 9 13 Mon 10 14 13 Ea Y Plan of Work 2 days Fri 10 11 13 Non 10 14 13 Y sunnary 1 day Tue 10 15 13 Tue 10 15 13 a Y Second report 16 days Wed 10 16 13 Wed 11 6 13 Ty Y Part 1 9 days Wed 10 16 13 Sat 10 26 13 Sv Y Interaction Diagram 4 days Wed 10 16 13 Mon 10 21 13 El Y Part 2 7 days Tue 10 22 13 Wed 10 30 13 y Z Class Diagram and Interface 4 days Tue 10 22 13 Fri 10 25 13 Ca Specification Bi Syste Architecture and System 4 days Fri 10 25 13 Wed 10 30 13 coi esign Figure 39 Plan of Work 1 89 O Task Nane 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 31 38 39 40 41 42 43 44 Part 3 Algorithms and Data Structures User Interface Design and Implementation Design of Tests Project Management and Plan of Work design info collection Twitter usage Crawling Twitter Data Storing Twitter Data Analyzing Twitter Data Visualizing Twitter Data First deno Report 3 analysis former report all reports collated summary Design 2 Network information Force directed layout Visualizing Temporal Information Visualizing Geo Spatial Information Visualizing Textual Information suggestion supply Second deno Electronic Project Archive Duration 5 days 2 day
29. things like location their recent activity browsing history search history and social networks Account settings page This page is where the user manages their personal account such as changing their login or password or location et cetera A sketch is listed below that explain most of what this section has enumerated 16 Figure 4 Paper prototype of the user interface 17 6 1 6 2 Functional Requirements Specifications Stakeholders Fitness Buffs These are considered the most serious of people in the system they are people who spend an inordinate amount of their time on managing their health and well being typically working out around 3 6 hours per day and managing their intake of vitamins calories complex sugars carbohydrates et cetera There is no other person in the system who is as serious about their fitness as they are Their interest is in their own fitness almost exclusively Average Consumers The average consumer is a person who does not work out regularly but is interested in the system and what it can help them do The system is geared towards helping this subset of person by encouraging them to work out and broadcast their activities while monitoring the community at large Their interest is in using the system to improve their health and in the health of his her community Business Owners Business owners are more narrowly people who own or operate gyms health food stores health
30. 5 Controller X x x x x Interface x x x x x Data Analysis x x x x Communicator x x x x x DB Connection x x x x x Table 8 Traceability Matrix 8 2 System Operation Contracts System Operation Contracts for the operations of the fully dressed user cases e Searching for Locations and Facilities 1 PRE CONDITION No precondition 2 POST CONDITION The user could achieve results they want e Viewing Statistical Breakdown 36 1 PRE CONDITION Data must be present in database 2 POST CONDITION The information is displayed to the user e Viewing Results 1 PRE CONDITION Data must be present in database 2 POST CONDITION The information is displayed to the user e Viewing Map 1 PRE CONDITION Data must be present in the database Google connection must be open 2 POST CONDITION The information is displayed to the user e Viewing Map 1 PRE CONDITION Data must be present in the database Twitter website must be available 2 POST CONDITION The information is displayed to the user 8 3 Mathematical Model 8 3 1 Before Gathering Data At any time the user will want to perform analysis on city or location theyre searching for The system will need to search for information based on tweets in the area of which there are N They will process data based on M hashtags identified b the system as being relevant to analysis There are assumed to be A users in a given area that tweet about health and B users in an ar
31. 9 System Architec 16 7 16 7 16 7 16 7 16 7 16 7 100 ture and System Design 15 points Sec 10 Algorithms and 16 7 16 7 16 7 16 7 16 7 16 7 100 Data Structures 4 points Sec 11 User Interface De 16 7 16 7 16 7 16 7 16 7 16 7 100 sign and Implementation 11 points Sec 12 Design of Tests 12 16 7 16 7 16 7 16 7 16 7 16 7 100 points Sec 13 History of Work 16 7 16 7 16 7 16 7 16 7 16 7 100 Current Status and Future Work 5 points Sec 14 References 5 16 7 16 7 16 7 16 7 16 7 16 7 100 points PROJECT MANAGE 16 7 16 7 16 7 16 7 16 7 16 7 100 MENT 17 points TOTALS points 33 4 33 4 33 4 33 4 33 4 33 4 200 Table 1 Contribution Breakdown Table 2 Contents 1 Individual Contribution Breakdown 2 2 Summary of Changes 6 3 Customer Statement of Requirements 7 Sal Problem Statement ld A eS re eee ei Soe nae Be ee 7 3 2 Background on Health Monitoring Analytics 9 3 3 Project Oyerview noos nde Miele RR AS ai celt ied deka 11 4 Glossary of Key Terms 13 5 System Requirement Analysis 14 5 1 Functional Requirements Table ex sais Pe dion ara EA A 14 5 2 Non Functional Requirements Table lt lt lt 14 5 3 On Screen Appearance Requirements 15 6 Functional Requirements Specifications 18 e DEE EE 18 02 Actors and COGS pa tas tb he Seas Oh he Sorat A
32. Mateus Santos Abdul Hassan evin Kobilinski Daihou Wang KBrien Range Sujana Gangadharbatla 93
33. Tweets This is a String variable that holds the Tweet query that is typed in by the user e Exercise This is a String variable that holds the exercise query that is typed in by the user Facility This is a String variable that holds the facility query that is typed in by the user The class methods are described below e setHotCities city Typical set function Is called when the user begins typing their query for a city and presses the search button The variable in the class is what gets set e getHotCities Typical get function Is used when the system needs to retrieve a Hot Cities from search results and send back to the screen e setFacility city Typical set function Is called when the user begins typing their query for a facility and presses the search button The variable in the class is what gets set e getFacility 65 Typical get function Is used when the system needs to retrieve facilities from search results and send back to the screen set Exercise exercise Typical set function Is called when the user begins typing their query for an exercise and presses the search button The variable in the class is what gets set getExercise Typical get function Is used when the system needs to retrieve exercises from search results and send back to the screen set Tweets Stweet Typical set function Is called when the user begins typing their query for a tw
34. an associated instance of the IdentityList object This would be called when the user is registering or trying to change their personal information as de tailed in the use cases Calling this function updates the value of the username variable and forwards the change to the database e getPassword 97 Typical get function Is used when the system needs to retrieve a password from an associated instance of the IdentityList object Returns the current value of the password variable set Password password Typical set function It is used when the system needs to change specify a pass word for an associated instance of the IdentityList object This would be called when the user is registering or trying to change their personal information as detailed in the use cases Calling this function updates the value of the password variable and forwards the change to the database getEmail Typical get function Is used when the system needs to retrieve an email from an associated instance of the IdentityList object Returns the current value of the email variable setEmail Typical set function It is used when the system needs to change specify an email for an associated instance of the IdentityList object This would be called when the user is registering or trying to change their personal information as detailed in the use cases Calling this function updates the value of the email variable and forwards the change
35. and their interaction in this software to be In this project we mainly have five classes and two derived class LocalUser and TwitterData Every time user reaches the user interface and makes a request then the controller will pass data retrieve request to the database In response database would send requested data package to the controller The controller would then call relevant functions with data package to meet the user request 10 2 Data Types and Operation Signature 10 2 1 Identity List The first class IdentityList contains all of the users personal information inside the following variables e Username String variable corresponding to the users username e Password String variable corresponding to users password e Email String variable corresponding to users email e Administrator Boolean variable indicating whether or not a user is an administrator TRUE or a normal user FALSE The functions of this class are listed below These include various set and get functions for class encapsulation along with functionality for the administrator to manage a given users account information e getUsername Typical get function Is used when the system needs to retrieve a username from an associated instance of the IdentityList object Returns the current value of the username variable e setUsername Susername Typical set function It is used when the system needs to change specify a user name for
36. arching for Locations Display Error message Database getTweet Location Display results map and chart Figure 6 Use Case Diagram la 25 B ei UC 1b Searching for Facility Input Facility Database getTweet Facility Process Data No data available for this facility WEE Display results map and chart Figure 7 Use Case Diagram 1b 26 H Buc 2 Viewing Statistical Breakdown Entering webpage Database getTweet All ate Retneve Data i Webpage Initiating Location categorization Tweets counting for each state Pie Chart Processing Display Pie Chart for Breakdown Figure 8 Use Case Diagram 2 27 N Me 3 Viewing Results Entering webpage Webpage Initiating Database getAllTweet Retneve Data Total Tweets Counting US Tweets Counting and calculation for estimated number of aio i Configuration of Wedges for SE Display Notification Buttons statistical Figure 9 Use Case Diagram 3 28 E ei UC 4 Viewing Map Entering webpage Webpage Initiating Database getAllTweet Retrieve Data Location categorization Tweet content extraction Marker map and Heat map KS i i tion Display Maps Figure 10 Use Case Diagram 4 29 A Buc 5 Viewing Word Cloud Entering webpage Webpage Initiating Database getAllTweet Retrieve Data lotal lags Extraction Tag Frequenc
37. ates the value of the sex variable and forwards the change to the database 60 getSex Typical get function Is used when the system needs to retrieve a sex from an associated instance of the object Returns the current value of the sex variable setNewName name Typical set function It is used when the system needs to change specify the name for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function updates the value of the username variable and forwards the change to the database getNewName Typical get function Is used when the system needs to retrieve a sex from an associated instance of the object Returns the current value of the username variable setNewPin pass Typical set function It is used when the system needs to change specify the password for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function updates the value of the password variable and forwards the change to the database getNewPin Typical get function Is used when the system needs to retrieve a password from an associated instance of the object Returns the current value of the password variable set Year year Typical set function It is used when the system needs to change spec
38. cture This is a First In First Out FIFO data structure that wed access through the use of pointers to objects in the list head and tail This isnt ideal for the storage because the operations can only be performed at the ends of the list and we may yet need objects in between e List This is highly similar to the Queue only that it is much easier to perform search ing addition and deletion operations e Map A map is a data structure where several variables of different types are stored in an abstract way and need to be referenced or pulled out according to a key pair value Addition removal modification and searching is quite easy with this data structure e Tree A tree is an abstract data type where nodes contain a series of values along with a pointer to the children of that node Its very fast to search for objects within the tree making it optimal for us Disadvantages are that nodes can only really access points that it has pointers to they cant reach the children of other nodes 74 e Vector This is a dynamic array that is random access and whose size is not fixed How ever it is not possible to insert at random points After considering the above data structures we decided to store the user information and characteristics in a Map data structure because there are several fields that are easy to get if you request the right key password name et cetera The Tweets we decided to store in a vec
39. d get a lot of noisy tweets not related to health at the same time Therefore we switch to some more practical ways We only pull the tweets having specific set of hashtags Although there is no any total number of tweets from one area for percentage calculation we can still compare the difference in health related tweet number across different states This also indicates the popularity of health and exercise across different states 38 9 Interaction Diagrams A Suc 1a Searching for Locations eh _ _ __ REESEN DR n m Input Location search getLocation string check with Google Senice Controller E oe DBConnection Database getTweet Location Process error mescana Display Error message Combined with Google Senice No data available for this location Controller AccessDisplay Controller re DisplayChart createMap oes Jee Map and chart Figure 14 Use Case a oe for Locations Use Case la is designed with the intention of giving the user the privilege of searching for location data through the user interface This is important for the general functionality of the application the ability to search for health related data is the crux of the design goals The interaction diagram displays the process the system needs to go through in order to handle a query from the user and correctly display the output The user needs to navigate to the index
40. de values are 0 for them Only a small portion of the tweets that we collected have location information provided with latitude and longitude Ideally we can use reverse geo coding to obtain the exact address for each tweet that has latitude and longitude values However Google map service for geo coding has special limitation of the number of requests we can make in one day Our massive twitter data needs more than Google s limitation Sticking with less cost policy for our class project we decided to use an alternative way the bounding box idea for each state Considering the difficulty to use bounding box for each city we divided the US tweets for different state To define a bounding box for one state we only need to know the minimum latitude minimum longitude maximum latitude and maximum longitude that the state covers Therefore we defined these four extreme values for the bounding box of each state All the tweet locations that fall into a box will be marked as tweets from that corresponding state Since bounding boxes are rectangular they wouldn t fit the state boundary perfectly This would cause some mistakes in the state determination However this is the second best way we can find besides using the reverse geo coding We obtain the extreme values of latitude and longitude for each state online Then all of these data are included in the program to estimate the state of each tweet location In other words each tweet with location
41. develop the system and improve its functionalities as well as maintain the website and interface Additionally they perform customer service and deal with maintaining the members quality of interaction 3 Web User Interface e Participating type 4 Tweets Database e Participating type 5 Google e Participating type 6 3 Use Cases 6 3 1 Casual Descriptions e UC 1 Searching for Locations and Facilities Requirements REQ1c REQ1la REQ11b The user wants to perform a search for locations or facilities nearby such that they can evaluate the statistics for that area for their own perusal They will be required to navigate to the search bar and enter a query and then press Enter e UC 2 Viewing Statistical Breakdown Requirements REQla REQ1b REQ3 REQ6 REQ7 The user wants to view the statistical breakdown for the geographical regions of the United States They will navigate over to the charts and graphs region of the website to view information such as tweet data broken down by location where the most frequent tweets are coming from and what time of day the most tweets are being generated e UC 3 Viewing Results Requirements REQ1b REQ2b REQ3 The user wants to view how many tweets are collected how many tweets come from the U S how many of them have a location and what fraction of the population that is tweeting is exercising 19 e UC 4 Viewing Map Requirements REQ1b REQ1c REQ2b REQ3 REQ4 REQ5 The user wants to know
42. e data the period collected initially was over a one month period starting near the end of October At the time we had believed that the Twitter filters applied by the API were working correctly and didnt check that the location filter and hashtag filter were properly working The MySQL database had been set up but when it was reviewed no outstanding errors were seen It was only when we begin to do the marker maps and filling in with tweet data text did we realize that the data was horrifically noisy we had gathered all tweets from the US and added onto that tweets filtered by our specific health related hashtags So the problem was that we did not understand specifically how Twitter applies its filters e g this was a programming error that caused the database to become full of junk tweets The database had to be flushed entirely And then the modifiers had to be played around with after that to make sure the tweets were being filtered correctly All told the demo represented tweets gathered over a period from 11 29 13 to 12 06 13 one week worth of tweets Twitter DOES allow you apply more than one filter to the data They have several different predicate parameters for querying the data To begin Twitter data is obtained by making an HTTP POST as I am sure you are familiar this is a standard HTTP protocol for web servers to make a request for data for the information We can select the presentation of the material XML or JSON We selected JSON b
43. e marker map contains clickable markers placed in tweets locations for displaying health related tweet content The interaction diagram shows how it gets implemented in our system This visualization requires the connection with Twitter database and the Google map service Following explains in detail about the data filtering and location categorization For the data filtering we only focused on the track parameter We only really had two options as far as we could see for searching for data e Filter by location and filter by hashtags This means we could pull ALL OF THE TWEETS from the United States along with ALL OF THE TWEETS worldwide 52 about healthcare The problem with this is were collecting huge amounts of junk data theres millions of tweets per hour in the United States The percentage of those that would be related to health must be astronomically low e Filter only by hashtag This means we would pull tweets worldwide about relevant subject material but we wouldnt be able to restrict it by location As such we would end up with tweets that have their GPS coordinates hidden as adjusted by the users privacy settings along with tweets that are outside the United States This seemed like the best solution because it would reduce the number of tweets and the noise while still keeping the material relevant As it would turn out only a small percentage of the tweets we pulled would be relevant to the US since the majority of them
44. e on Software Engineering ICSE 2103 Honolulu Hawaii May 2011 Great deal of useful information and relevant papers is available on the site by Frank Bentley Yahoo Health Mashups Ming Hao Christian Rohrdantz Halldr Janetzko Umeshwar Dayal Daniel A Keim Lars Erik Haug Mei Chun Hsu Visual Sentiment Analysis on TwitterData Streams Mateus Santos Abdul Hassan evin Kobilinski Daihou Wang KBrien Range Sujana Gangadharbatla Workout with Friends Health Monitoring for Fitness Applications Web user Interface http creately com diagram example goc8uhkx Copy of Web iRis User Interface Design http www elsevierdirect com companions 9780120884360 casestudies Chapter_01 pdf Methods for improvement http www doc ic ac uk nd surprise_97 journal vol2 hafj Methods to create a user interface user experience mock ups http www codingrobots com screensketcher Create Mock ups and wireframes http spyrestudios com 15 quality web based applications to create mock ups and wireframes Estimating with Use Case Points http www cs cmu edu jhm Readings Cohn Expert estimation http www idi ntnu no grupper su publ ebse RK15 reviewexpertestim jorgensen jss04 pdf Compared Formal Models and Expert Judgment http citeseerx ist psu edu viewdoc download doi 10 1 1 163 9404 amp rep rep1 amp type pdf Genetic programming GP http profs info uaic ro ogh files sbse articles new articles sdarticle pdf 92 26
45. e tiers are based on the deployment of layers scaling out an application is reasonably straightforward e Flexibility Because each tier can be managed or scaled independently flexibility is increased e Availability Applications can exploit the modular architecture of enabling systems using easily scalable components which increases availability 11 2 Identifying Subsystems Web Browser Client N Web Server Serve N i Network Server Logic HTML Server Resource D D DataController Communication JavaScript y Layout amp Display TwitterDaraBaze Figure 25 Logic Diagram of the VGA Controller Block In this system there are two subsystems One is the client side the web browser the other side is server web server The client package mainly refers to the website framework and it contains the building structure of the user interface For the server side when the user uses sends a request to the web server the server would response through processing it in its logic package The logic unit would analyze it and further pass it to the server resource package This package is responsible for retrieving data from local Twitter database and internal communication 11 3 Mapping Subsystems to Hardware As shown in Figure 4 1 the subsystem can be mapped into the following hardware component easily The Web Browser package are allocated in c
46. ea that do not tweet about health and are for simplicity assumed not health For simple statistics the analysis becomes Of Healthy Population Population Census 1 A B 8 3 2 After Gathering Data There is always noise present in the data when we search for tweets and the system needs to compensate for that There will be users who discuss information in a sarcastic way or attach hashtags to their tweets that are irrelevant to what theyre saying and can be mislabeled by the system To compensate we need to target a users tweet directly and then search their history If their history gives an indication that they are health conscious or that theyve used this hashtag multiple times before in a relevant way gauged by retweets then the data is compensated As time goes on tweeters will be given a probability weight that their tweet is useful and compared to a threshold We collected all the tweets about health and exercise with a set of hashtags After filtering out all the tweets having no information about location we counted the number of tweets in US concerning about the health and exercise noted as B After researching online we 37 found that 8 of US people use Twitter Therefore dividing the number of tweets from US by 8 denoted as B 0 08 we get the number of people in US who exercise But this is not the total number of people in US who exercise because we havent collected all the health related tweets in US T
47. eb browser Chrome Firefox Internet Explorer etc rather than installing the application in a Client Server structure This greatly simpli fies the client computers load reduces system maintenance and reduces upgrade costs and the workload and the overall cost of the user TCO It is differentiated from C S structures 2 tier architecture B S structure usually needs a middleware between the front end and the backend forming a 3 Tier Structure 11 1 2 Tier Architecture This is well suited to separate the roles of three components of the system the presentation tier the application tier and the data tier And each component is a tier that can be located on a physically separate computer They generally use platform specific methods for communication instead of a message based approach Different from 2 tier model the 3 tier architecture describes the separation of functionality into segments which independently keeps the process management and model logic delineated Additionally 3 tier architecture allows any one of the three tiers to be upgraded or replaced independently 69 Compared with the 2 tier B S model this 3 tier architecture style will improve system reliability and flexibility The main benefits of the 3 tier architectural style are e Maintainability Because each tier is independent of the other tiers updates or changes can be carried out without affecting the application as a whole e Scalability Becaus
48. ecause were using Javascript on the 45 server side to data processing and rendering The data when queried from our DB would be sent as a JSON to be parsed by Javascript later on As for the query structure an example would be as the one they supply in their API code https stream twitter com 1 1 statuses filter json delimited length amp track twitterapi This is what we are POSTing to Everything after the filter json are parameters that we can set ourselves and then append onto the URL with amp symbols The parameters are e Follow Follow is a comma separated set of user IDs So follow twitterapi to find twit terapi tweets The server will return all tweets for the specified user IDs con taining x User generated tweets Tweets retweeted by the user Replies to any tweet the user made Retweets of the users tweets by other people It does not have Tweets with the users name in it x Tweets from users with privacy centric settings e g private users e Track These are the list of keywords to track in other words the hashtags This is also a comma separated list of characters We are not able to do exact search ing e g like typing donuts krispy Kreme into Google with quotes to ensure the results contain these words We can t exactly delimit by hashtag either because the streaming API for this parameter when we search for twitter returns any instance of it TWITTER twitter
49. ecific topics in natural language or by a specific hashtag People can find where the healthcare conversations are taking place discover who to follow within your specialty or disease and find the best from conferences in real time or in archive Unfortunately this is more focused on an academic research type of community It is used to globalize and clarify health care specific topics diseases et al HMS Health Monitoring Systems The EpiCenter system is capable of analyzing healthcare data for the purpose of de tecting anomalies suggestive of public health threats such as disease outbreaks and bioterrorism Users can find reported data based on the location settings and other options However the system is not able to reflect the actual condition of the com munity which we know is changing in real time and is not specific this is a macro scale application of analyzing the health of a community through viral outbreaks and is meant for hospitals and not communities 10 Figure 3 HMS Health Monitoring Systems These are the only two products we could find being offered to consumers and neither of them even combined fully implement what we plan to do here So we can conclude with some aplomb that this approach is highly innovative 3 3 Project Overview To fulfill the requirements of our customer statement we have defined the following criteria as being mandatory for implementation with the software 1 Generate rea
50. eet and presses the search button The variable in the class is what gets set get Tweet Typical get function Is used when the system needs to retrieve tweets from search results and send back to the screen 10 2 8 SendMessage This class implements the functionality described in previous use cases for when users want to communicate directly with each other through the website The class variables are as below e userID A String variable containing the userID the user wants to message e Message A String variable containing the message the user wants to send to the user specified by userID The class methods are explained below e findFriend userID This function will loop through the user database until it finds a user that matches the userID effectively a series of get calls e addFriend userID Takes the userID and updates the String pointer in a users database to include this new ID as well e deleteFriend userID 66 Takes the userID and updates the String pointed in a users database to remove the ID associated with the variable e writeMessage Smessage It allows user to edit text content he she wants to send as a message to friends If the content is legal it returns 1 and would calls the function messageFriend userID otherwise it returns 0 and requests a re enter to user e messageFriend userID It works as a communicator in this class It is called by wr
51. er side needs a server computer for providing local Twitter database The computer can run basic database as well as Java IDE It should have enough disk storage for example 4 Gbytes hard disk space and be able to perform network communication For the client side it needs a normal computer that can open websites and access to the Internet 12 Algorithms and Data Structures 12 1 Algorithm 12 1 1 Estimation of population concerning health Originally we were planning to get all the tweets from one area either a city or a state Then we planned to find out the number of tweets having physical exercise information by finding some keywords from the tweet content Then we would have the total number of tweets T and the number of tweets that are health related A In that case we can calculate the percentage of people in that area who concern about health and exercise as shown in following formula A Of Healthy Population T Population Census 2 As we start implementing it we realized that this is impractical The number of tweets from a city for example NYC is so large and the number of tweets mentioning health and exercise is so small We have storage problem for the massive twitter data and we would get a lot of noisy tweets not related to health at the same time Therefore we switch to 71 some more practical ways We only pull the tweets having specific set of hashtags Although there is no any total number o
52. es us another approach to analyze the data We divided the tweets into four divisions based on their location northeast midwest south and west We processed the data for each location division For example we need to analyze the tweets from northeast In the program we have a loop to go through each tweet For each tweet we extract the time string to determine which hour the tweet is posted Then we count the number of tweets for each hour We plotted tweet count vs hour in day curves for four US regions The curves change when we have larger amount of tweets We constantly pulling data from twitter and our website automatically updates with the most recent data From the curves we showed in demo 2 west seemed to tweet about exercise at about 6pm after work Figure 2 shows the current distribution of tweets over 24 hours of day People from west blue curve tweet more about exercise in 7pm judging from the curve peak at 17th hour of days As we can see from current curves people in south tweets most at around 10am and 5pm judging from the two peaks of the curve The curves generally becomes more flat comparing to what we had in the demo day We can still see the hours when people tweet more from the peaks of curves The red curve shows the total US people tweeting activity People tweet more during the daytime from 9am to 10pm than other time periods such as 12am to 6am wall Statistics sw CS Figure 19 Statistical Graph 50 It i
53. est D Data Analysis Render search request to Tweet Database D Communicator Access Tweet Database and execute request D DB Connection Receive notification from Data Connection D Controller Table 5 Responsibilities Types and Concepts Table 8 1 2 Association Definitions Some of the concepts defined above as domain concepts have to work in certain patterns to finish some target requirements Table 5 2 gives the corresponding association definitions based on the defined domain concepts 35 Concept Pair Association Description Association Name Database Connection Database forms a DBConnection Store Retrieve DB to send information to the system Data Analysis DBConnection passes the information from Render DB Connection the Database to DataAnalysis Database Database sends information through Calculate Data Analysis Data Analysis to be processed Table 6 Association Definitions 8 1 3 Attribute Definitions Responsibilities Attribute Concept Know if the user input keywords to search Awaiting Search Request Controller for data and information Display a map to show results Map Interface Display a list of results that have been searched Search Result Interface Display analyzed data in visualized form Data Display Interface Analyze data from Twitter database Analyze Data Data Analysis Table 7 Attribute Definitions 8 1 4 Traceability Matrix Domain Concept UC1 UC2 UC3 UC4 UC
54. ey should be kept e UC 2 Selecting Hot Cities This is another user or member input option as selecting hot cities instead of searching through search bar The test should make sure the system gets what the user selected and make sure that the system has a good user interface to avoid any possible mistakes due to in appropriate operation e UC 5 Accessing Personal Page This use case tests functionality of the personal page There are several processes that occur when its being accessed The test should make sure the system gives full consideration on whether the user is registered before logging in the verification of the identity whether the searching algorithm for this user s page is efficient and whether the interface of the personal page has any flaw 83 UC 6 Editing Personal Information This use case test should focus on the UI and data connection with member database The test should make sure whether the information is successfully updated if the logged in user member has changed his information A lot of emphasis should also be put on the user interface testing to make sure the system functions correctly UC 7 Sending Messages In order to send messages the user has to be a member of our software system The test should make sure the sender is a registered user after sending message and at the same time the receiver should be notified The message should be sent without error The system should also take into account po
55. f the issues relating to why people dont exercise e For many of you no time is a baseless complaint it often arises among those who are unable to manage time effectively themselves Obtaining a map of someones daily using tweets can allow us to if effectively compared against a person who has similar time constraints that is exercising effectively generate a plan of action or socially network the lazy man and the exercising man together e We should be able to provide a means to understand why a person has no energy from their social media data if there is any A mechanism should be put in place to provide suggestions on how to improve their energy for exercise be it comparison to their friends small changes in diets using advertising or adjusting their schedule to find hours of the day where they dont feel tired and suggesting exercise there e For the problem of competing interests a person could be informed about a type of opportunity cost related to doing one task versus not exercising at all e This part is easy enough it is only necessary for the person to be informed about how their friends might be changing their activities or by viewing the status of a community health board The last issue to address is of a personal nature to you the consumer We want the software to be free and readily available to you at no cost We expect it to be easy to find what type of information you are looking for We want you to be able to use w
56. f tweets from one area for percentage calculation we can still compare the difference in health related tweet number across different states This also indicates the popularity of health and exercise across different states We collected all the tweets about health and exercise with a set of hashtags After filtering out all the tweets having no information about location we counted the number of tweets in US concerning about the health and exercise noted as B After researching online we found that 8 of US people use twitter Therefore dividing the number of tweets from US by 8 denoted as B 0 08 we get the number of people in US who exercise E B Re 3 But this is not the total number of people in US who exercise because we havent collected all the health related tweets in US The tweets we obtained is just a subset of all health related tweets from US As we collect more data the number of collected tweets will increase so as the number of people in US who exercise Our estimated number will get closer and closer to the true value as time goes There are some limitations on the accuracy of our estimated number of people in the US who exercise We can only get closer to the true value by may never be able to reach that number Another factor is that there may be more than one tweets about exercise sent out by one person This person is counted more than once Future improvement on this project can refine the algorithm to count the number
57. he tweets we obtained is just a subset of all health related tweets from US As we collect more data the number of collected tweets will increase so as the number of people in US who exercise Our estimated number will get closer and closer to the true value as time goes There are some limitations on the accuracy of our estimated number of people in the US who exercise We can only get closer to the true value by may never be able to reach that number Another factor is that there may be more than one tweets about exercise sent out by one person This person is counted more than once Future improvement on this project can refine the algorithm to count the number of Twitter users who tweets about health and exercise instead of the number of tweets Originally we were planning to get all the tweets from one area either a city or a state Then we planned to find out the number of tweets having physical exercise information by finding some keywords from the tweet content Then we will have the total number of tweets T and the number of tweets that are health or exercise concerned A In that case we can calculate the percentage of people in that area who concern about health and exercise As we start implementing it we realized that this is kind of impractical The number of tweets from a city for example NYC is so large and the number of tweets mentioning health and exercise is so small We have storage problem for the massive Twitter data and we woul
58. herever you want mobile or desktop We also want to implement privacy options so that you have total control over what you are showing to your friends or to your community but we hope you will share as much as possible It takes a village to make a man 3 2 Background on Health Monitoring Analytics Most systems today for personal activity monitoring focus on the benefits and rewards for an individual user The user may share this information with his or her friends but everything revolves around an individual We have aggregated some active research work in analyzing the health of an entire community based on big data from a social network 1 Healthcare Hashtag Project symplur WHAT CONTACT The Healthcare Hashtag Project HEALTHCARE HASHTACS REGULAR TWEET CHATS CONFERENCES DISEASES Why the Healthcare Hashtag Project The Community by the Numbers PLACE DISCOVER WHO TO FOLLOW WITHIN YOUR SPECIALTY OR DISEASE Ll 43 2 949 mma 9 DISCOVER THE BEST FROM CONFERENCES IN REAL TIME OR IN ARCHIVE 1 203 Kigi Figure 2 The front page of the Healthcare Hashtag Project The stated goal of this project is to make the use of Twitter more accessible for providers and the healthcare community as a whole How it works is that users are able to use a search bar to scan the web for relevant hashtag data The web would then extract data from Twitter and display it in an analyzed digestible form for the user Users can it to search for sp
59. his particular city In addition the output window also shows some notifications depending on how this citys health awareness comparing to other cities in our database For example if the city shows low percentage of people mentioned about physical activities and health concerns the system will output a notification saying According to our analysis this city has relatively low health activities comparing to others People in this city are encouraged to exercise more Or some notification about how it is changing with time For example The health activity is decreasing recently in this city The data analysis window will show the variation of this statistical result in percentage corresponding to time The visitor can see how peoples awareness concern of their health changes 31 over time in this city It can also switch to other types of diagrams such as tables when the visitor click on the switching layout within this graphic window Alternatively the visitor can click on one of the hot cities shown under main page section Same results will show in these output windows as stated above However in this case visitor doesnt have to input the city in the search bar The visitor can input a health facility in the search bar The map will mark all the locations of the health facility in the United States The visitor can zoom in to different cities to see the distribution of these facilities in different places For example searching fo
60. hould make sure the map can be easily manipulated The user should have an option to request the display of partial data if the data set is too big However the user can still obtain full statistical results from the output window 84 e UC 14 Viewing Word Cloud This is a dynamic graphic on the key word from twitter and can be displayed in a globular figure it can change size and position with mouse and connected to detailed information with clicking e UC 15 Viewing Line Chart This is a Line chart to display the average sport time in different place it updated over time and implies local lifestyle and sport degree e UC 16 Viewing fan chart This is a fan chart to displace sport degree in every single state and for other appointed region and also display the ratio We also display some dynamic animation effect on move cursor in each section 14 2 Test Coverage e UC 1 Searching for Locations and Facilities Success User inputs a location they want to query The system undergoes the task of verifying if the location is valid and can be displayed When the match occurs the MapFragment is updated along with other relevant portions of the user interface User inputs a facility they want to query The system undergoes the task of verifying if the facility is valid and can be displayed When the match occurs the MapFragment is updated along with other relevant portions of the user interface User inputs a friend they wa
61. how the tweets are distributed graphically using a map The user can view the information as a distribution via a heatmap or directly view all of the tweets plotted on a map where the marker can be interacted with to show the content of a tweet e UC 5 Viewing Word Cloud Requirements REQ2a REQ3 REQ9 REQ10 A user wants to view a word cloud containing the hashtags used to filter the data They will visit the website and view the word cloud from their home page which could reflect the frequency of the hashtags depending on the size of the tag inside of the cloud Clicking on the tag will take the user to Twitter 6 3 2 Use Case Diagrams The use case diagram is shown in the figure below Member Visitor and Administrator initiate all use cases except for UC 2 Choose Hot Cities which is extend from UC 1 Location facilities Search as sub use case Database A store tweet data for Member and Visitor Database B store members information for Member and Administrator Thus they are participate in all use cases Member is generated from a Visitor when the Visitor chose to run UC 10 Register 20 Use Case Diagram Case Diagram ei lt lt Participate gt gt Figure 5 Use Case Diagram 6 3 3 Fully Dressed Descriptions e UC 1 Searching for Locations and Facilities Initiating Actor Member Actor Goal To search available location or facilities Requirement Addressed REQ1c REQ10a REQ10b PRE CONDITION
62. ick done to go back to the main page one click e Click on drawdown button and select personal page button two clicks e In the personal page look for a friend in the recommended list and click symbol to add one click e Done adding a friend e The page will show the access to this friend The member who wants to send a message to a friend e Click log in button one click 33 Input user name and password in the coming page several keystrokes Click done to go back to the main page one click Click on drawdown button and select personal page button two clicks In the personal page select one added friend one click A message box shows up asking for message content Type in message several keystrokes 8 Click send one click 34 8 User Interface Analysis 8 1 Domain Model Database Store Retrieve dBnetworkAddress Ca ate Render DataAnalysis MapDi spiay User Resul Display Communicator IsWait ingSearchRequest Figure 13 User Domain Model 8 1 1 Concept Definitions To analyze the domain model we first derive the domain model concepts and corresponding responsibilities from the formerly defined system use cases Table 5 1 lists all the domain model concepts and corresponding responsibilities Responsibilities Type Concept Handle requests from user Controller Display data in numerical graphical form Interface Analyze data depending on the search requ
63. ify the year for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function updates the value of the year variable and forwards the change to the database set Month month Typical set function It is used when the system needs to change specify the month for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function updates the value of the month variable and forwards the change to the database 61 setDay day Typical set function It is used when the system needs to change specify the day for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function updates the value of the day variable and forwards the change to the database get Year Typical get function Is used when the system needs to retrieve a year from an associated instance of the object Returns the current value of the year variable getMonth Typical get function Is used when the system needs to retrieve a month from an associated instance of the object Returns the current value of the month variable getDay Typical get function Is used when the system needs to retrieve a day from an associated instance
64. ing to users password Email String variable corresponding to users email e Age Integer variable corresponding to the users age e Year Integer variable corresponding to the year the user was born Month Integer variable corresponding to the month the user was born e Sex Boolean variable corresponding to gender FALSE Female and TRUE Male e Day Integer variable corresponding to the day the user was born this is day of the month not day of the year The functions for this class are outlined below e setAge Sage Typical set function It is used when the system needs to change specify the age for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases or when the system is attempting automatic updates once the users birthday has been reached Calling this function updates the value of the age variable and forwards the change to the database e getAge Typical get function Is used when the system needs to retrieve an age from an associated instance of the object Returns the current value of the age variable e setSex sex Typical set function It is used when the system needs to change specify the sex for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function upd
65. iteMessage message functions and sends the edited message to the target user ID It also utilizes the findFriend userID function to locate the correct user ID 10 2 9 Controller This class works as a central controller and communicator in the system to be It con tains seven derived functions DBConnection CheckIdentity CreatePersonalProfile UpdatePersonalProfile AccessPersonalProfile MessageManager and AccessDisplay Each function is linked to the other classes to perform user requests It also contains ten status of the current system This class is essential to system performance we would get nowhere without it e DBConnection It establishes a connection to the database for either stored local user profile or Tweet data e CheckIdentity It links to the IdentityList class when users try to login or edit profile CreatePersonalProfile It links to the AccessProfile class to create personal profile when a new user registers for the system It also links to Database class to store the newly created data UpdatePersonalProfile It provides a link to both of the AccessProfile and Database class when the logged in user attempts to edit his her profile AccessPersonalProfile It links to the AccessProfile class for users to gain access to their profile MessageManager 67 It links to the SendMessage class to send messages between different users e AccessDisplay
66. ited States rather than specific communities As such we have struck out this concept of a hot city and instead focus on state level segregation of tweet data We streamlined our use cases down into the few that the system now implements There are clarifications added to specific use cases that deal with issues on using the software The algorithms section has been updated to match our statistical analyses and useful equations when dealing with the weight of the word cloud There are discussions on filtering that can be seen in the section on Interaction Dia grams Several sections have been updated in the sense that they add more descriptions on the final product as well as just cut down on material that is no longer relevant and this applies to most if not all sections including the Customer Statement of Requirements 3 Customer Statement of Requirements 3 1 Problem Statement To our devoted consumer base It is our belief that the most important factors to a happy and healthy life is physical exercise and a balanced diet We can work our fingers to the bone and party to our limits but it isnt possible to live the most balanced life without first balancing our health Some of you out there struggle with a variety of conditions related to your health type 2 diabetes high blood pressure high cholesterol et cetera It would be much easier to manage these illnesses or even eradicate some of them if you could work up the wi
67. ks the available city and updates the statistical data accordingly Failure User selects a hot city but the system is unable to verify it It displays a No Results Found error message and waits for new input from the user e UC 6 Editing Personal Information Success The user will opt to change the information stored in the database that correlates to their personal profile They can change their email location gender age et cetera If done correctly the database will be updated with the new values and the screen will refresh to show that Failure x The server is unable to change their location An error message is displayed on the screen and requests that the user try again x The server is unable to change their gender An error message is displayed on the screen and requests that the user try again x The server is unable to change the age An error message is displayed on the screen and requests that the user try again x The website is unable to communicate with the server An error message is displayed on the screen and requests that the user check their connection or try again The website screen doesnt refresh to show the updated information 86 The website screen refreshes and the data is changed in the database but the information displayed remains unchanged e UC 7 Editing Personal Information Success The user will select another user they wish to send a message to If executed co
68. l time health statistics for a certain area We expect to be able to aggregate the Twitter data and extrapolate it to figure out how much of a community or city is physically active and how much it isnt We also want to know the distribution of the involvement in different regions This should be updated as soon as tweets involving exercise come in 2 Heat Map Based on the existing tweets from Twitter we developed a graphical representation of the distribution and concentration of the tweets across a given area a heat map where different colors indicate different intensity levels And how much of a community or city is physically active in sports or other health related topics 3 Marker Map We will also display an output of the exact tweet data in a location on a map that 11 is separate from the heat map This gives us the more discrete nature of the tweets versus the distribution although it can be used for that as well Tag Cloud Tag cloud is used to display the frequency of the hashtags used when querying Twitter for tweets The bigger the size of a hashtag is the higher its frequency is When clicking a hashtag the user will be redirected to the Twitter search site that shows the most recent results for that tag 12 A Glossary of Key Terms Hashtags A word or a phrase prefixed with a pound symbol Its typically used in social media for denoting important phrases On Twitter its used for grouping topics
69. land 0 00 EB Connecticut 11 46 Ei New York 21 88 IEI Pennsylvania 7 84 EB New Jersey 45 23 FOX US West Region El Maryiand 8 51 E Virginia 1 78 IE North Carolina 7 72 EI Georgia 8 71 IEI Fiorida 23 76 E Kentucky 4 75 EI Tennessee 3 56 IEI Mississippi 0 50 IEI Alabama 4 55 IEI Oklahoma 5 25 Arkansas 277 E Louisiana 2 38 Figure 28 Whole UI E Wisconsin 0 00 E Michigan 26 72 IEI nois 27 10 IEI indiana 9 92 EI Ohio 3 02 IE Missouri 5 34 E North Dakota 0 38 IEI South Dakota 0 78 EB Nebraska 0 70 EI Kansas 5 73 ES Minnesota 8 54 lowa 5 73 ES idaho 271 Fi Montana 0 00 IEI Wyoming 0 24 ED Nevada 0 00 EI utah 1 50 IEI Colorado 3 73 E Arizona 5 42 EB New Mexico 1 60 IEI Alaska 0 00 El Washington 8 14 E Oregon 5 42 California 70 85 E Hawaii 0 00 In our Ul part we kept the basic structure but modified the functionality We removed the hyperlinks from the dashboard and left all of the functions on the main page We designed several data analysis options including a pie chart breakdown of tweets a heatmap for the distribution markers on the maps and a graph of tweet volume versus time of day aggregated over a period The user effort with this screen is quite small The webpage is clean and simple it is ea
70. lients PC and the user would use this to access the application user interface The Web Server as well as Server Logic package is allocated in the server They would use this to appropriately analyze and process the user requests The Server Resource package is allocated in the central server 70 11 4 Network Protocols There are many network protocols such as FTP File Transfer Protocol HTTP Hyper Text Transfer Protocol and SSH Secure Shell In this system a user uses the website to log into his her personal account and access the software user interface Since we have a web design we need to use HTTP so users can navigate to the website from their personal computer 11 5 Global Control Flow For execution order in this software to be it is event driven so the system is waiting in a loop for emerged events user requests For example an user may want to search for the hot city New York City in the search bar before this request the system is waiting in a certain a loop for interruption after receiving the interruption the system jumps out of the waiting loop and executes appropriate subroutines The designed system to be is a real time system which means the system would continuously download tweets hashtag as well as user profiles from Twitter and store valid information in the local Twitter database This system to be is not a multiple threads application 11 6 Hardware Requirements To achieve full functionality the serv
71. ll to exercise It is often too late for many of you to realize that theres something wrong until your body hits a point of peak physical distress communicating through you with the use of pain Pain is an undesired signal because e You do not enjoy pain e It arises in moments of severe distress i e when something has gone wrong enough to the point of medical attention e It often comes too late when you d want to anticipate the problems pain is alerting you to in advance We here at our organization understand the various problems related with exercise In a poll of our customers the most likely reasons why you don t exercise can be found on this list e No time e No energy e Competing interests e Haven t developed the habits e No motivation e Too overwhelming e Poor diet e Current physical condition e No access e Lack of results So how can we work together you the consumer and us the company to break down these barriers to entry for a healthier life We have heard and understand your complaints and well take this opportunity now to explain how we think we can help you get in shape The operating principle is this man is a social creature it does not do well on its own 7 and needs community support to survive and thrive This logic can be applied to you the consumer as well you are more likely to perform a task if you know you will have a compatriot along for the ride As such we know you can be motivated to
72. ms the front end and it can be accessed by any computer Information is displayed in a static form and the user isnt given any ability to directly modify the Database except by initiating process management functions e g changing password e Process Management This middle level forms the core functionality behind the user interface It controls application functionality by performing detailed processing with related function calls as seen in the use cases This tier is concerned with system functioning and processing functionality e g searching for data recording updated information retrieving login information and related tasks e Database Database in this tier is kept independent of application servers and user interface The database management system on the database server contains the computer data storage logic When a request is sent to this tier database will look up the desired data Our system is a web based application which needs to serve a large amount of users Much of the functionality weve discussed so far depends on database interactions and queries We determined that a system like our software needs Browser Server architecture so that it can set up a communication pathway between the user side and the data side 11 1 1 Browser Server Structure and 2 Tier Architecture Browser Server B S structure is an extension of Client Server C S architecture In such a structure the user interface is through a w
73. ne click e Navigate to our software webpage several keystrokes inputting http address e Main interface page is brought to the visitor e Close our webpage when finished one click 2 DATA ENTRY several keystrokes and one click 32 e Input city facility in the search bar several keystrokes e Click on the search icon one click e The analyzed statistical data and graph will show to the visitor 0 effort 3 Or Alternatively only one click e Click on one of the hot cities in main page window one click e The analyzed statistical data and graph will show to the visitor 0 effort The visitor who wants to register to a member 1 REGISTRATION NAVIGATION two clicks e Click on log in button one click e A new page pops up asking for user name and password and an option of regis tration 0 effort e Click on the register link one click e A registration page pops up asking for information e Done with registration 2 INFORMATION FILLING several keystrokes e Account Registration Part 1 Instructions and how to use the application e Account Registration Part 2 Disclaimers and Permissions e Account Registration Part 3 User information e Done with registration and a personal page is set at the same time The member who wants to add a friend 1 ADDING FRIEND five clicks and several keystrokes e Click log in button one click e Input user name and password in the coming page several keystrokes e Cl
74. nessmotivation settingfit dreambody fitnessgoal resultssofar transformation goal dontgiveup weightloss morningrun workoutsoicaneat TFLers tflers gymlife trainhard grow bulk squat squats instafitness ripped swole muscle shredded pushup pushups sweat erind fitnessgear pushpullgrind bigbench fitgram pullup pullups gymrats eattogrow physique nikeplus nike fitnessmodel fitnessaddict trainer yoga ironaddicts lifting addictedtoiron bootcamp workout_routines strengthtraining smashfit squash swimming biking cycling losing weight keepfit partial SELECT COUNT FROM tweet_tags WHERE tag count 0 for index 0 index count tags index query partial query query tags index result mysql query query value mysql_fetch_row result Sfreq Sindex value 0 So we are explicitly searching for our hashtags because we have many many hashtags in the database were interested in seeing how well OUR tags are performing not all the tags we have just yet We can obtain how many times they occur in the tables this is the freq variable Then on the data weighting size we determine the tag size heuristically b
75. nt to search for The system undergoes the task of verifying if the friend is valid and if a list of matches can be shown on the screen Failure User inputs a location but the system is unable to verify it It displays a No Results Found error message and waits for new input from the user User inputs a facility but the system is unable to verify it It displays a No Results Found error message and waits for new input from the user e UC 2 Selecting Hot Cities Success User selects a hot city that is available from a list on the screen The system checks the available city and updates the statistical data accordingly Failure 85 User selects a hot city but the system is unable to verify it It displays a No Results Found error message and waits for new input from the user e UC 2 3 UC 4 UC 5 Viewing Graphical Data Results Map Word Cloud Success The object under question in the above list delimited by will correctly display to the screen if the server is able to interface with the website and update the information content on the screen Failure The website is unable to communicate with the server An error message is displayed on the screen and requests that the user check their connection or try again 14 2 1 Deprecated Cases e UC 2 Selecting Hot Cities Success User selects a hot city that is available from a list on the screen The system chec
76. of a hashtag comparing to others are reflected with the size of the word in the word cloud when it is in focus The interaction diagram shows the general data process of word cloud implementation Following shows the hashtags for the word cloud 54 fru running gym 08 jogging jogger marsic exercise exercising FFexercises fitness HRE Hitfam Ffseenonmyrun fitnessfriday fitforlife 4 mefirst Feymtime healthy bodybuilding chest workout H eetfit active healthychoices diet cardio Fstrong Ftraining fitnessaddict HHitspo instahealth Ftrailrunning trailrunner Ftrailrun Hitspiration fitnessmotivation gettingfit Fdreambody Hitnessgoal Fresultssofar Fransformation goal dontgiveup weightloss morningrun workoutsoicaneat HTF Lers Fflers Feymilife Ftrainhard Ferow Fbulk HFsquat HFsquats instafitness ripped FFswole muscle FFshredded Fpushup Fpushups Hsweat HFerind fitnessgear Fpushpullerind FFbigbench Hitgram pullup pullups Feymrats Feattogrow Fphysique Hnikeplus Hnike Hitnessmodel Hiitnessaddict trainer Hyoga FHronaddicts lifting addictedtoiron bootcamp
77. of the object Returns the current value of the day variable setEmail Semail Typical set function It is used when the system needs to change specify the email for an associated instance of the class This would be called when the user trying to change their personal information as detailed in the use cases Calling this function updates the value of the email variable and forwards the change to the database getEmail Typical get function Is used when the system needs to retrieve an email from an associated instance of the object Returns the current value of the email variable manageFriendList This operation allows a user to manage his own friend list It prompts the user with his her friend list and gives the user the freedom to add or delete a friend managelnbox This operation prompts the users inbox and shows his messages A user can delete or save messages changeProfilePicture This would allow a user to change his picture on his her personal profile 62 10 2 4 Database The Database class defines operations connected with database This class is in charge of storing and updating users personal information as well as getting data for other classes There are 5 operations and 2 derived class e storeNewName name This function would be called internally within any setName function and would update an associated value e g username within in the DB for that update
78. or by finding the option on the landing page if they are not already users A registered users information is entered into the Member Database 23 e UC 11 Backup Member Data An administrative user wants to back up database memory in the event of runtime failure They will requisition the system to store data in parallel one set is actively updated during system runtime and the other set is updated periodically by polling the active set e UC 12 Deleting Member Account A user wants to delete their account because they are either not using the service anymore have made a new account or are taking a leave of absence from it They will have the option of permanent deletion from the member database versus a suspension of the account where it can be reactivated but is no longer visible by other users 6 4 Traceability Matrix UC We 03 1 49 Req la x 1b x x x x lc x x 2a x 2b x x 3 X X x x 4 x 5 x 6 x 7 x 8 x 9 10a x 10b x Table 4 Traceability Matrix 6 5 System Sequence Diagrams The traceability matrix shows the distribution of the requirements for each of the use cases These use cases are the key features of our software Any user of our software can access these The user will interact with the the system the website and from there the websitre will interact directly with the database and pipeline results back to the user 24 Al ei UC 1a Se
79. products lines et cetera Their interest is in using the system to figure out how to target consumers that are more physically active and thus more likely to use the product or to find new venues in areas that are lacking them but have the potential to support them Academics Researchers This set includes those who work in universities or in corporate research that are working in fields that require aggregation and analysis of data related to the health of communities The members could be from many different disciplines nutrition dietetics psychology sociology et cetera Their main interest is in using the system to further research results Government Officials Government officials is a highly diverse and wide spanning group that includes people from local townships and municipalities all the way to the state and federal levels with job titles ranging anywhere from councilperson to state senator Their interest in the system is in using it to guide policies or reform for a targeted region by changing laws that could have an impact on health e g Bloombergs soda ban in New York City Actors and Goals 1 User e Initiating type 18 e Goals To interact with the system to find information about the health of a region to find people in his or her area who exercises a lot and to leverage the systems full functionality to assist themselves in living a healthier life 2 Administrator e Initiating type e Goals To further
80. r swimming pools would result in showing all large popular swimming pool facilities in the United State At the same time the map will mark twitter users who mentioned similar topic in their tweets with blue dots and mark all twitter users in database with red dots The visitor can see distribution and percentage of people mentioning about this facility or the corresponding activity The member user who has registered can use these two features stated above that all visitors can use The member can log in to use some additional features about accessing his her own page managing friends and sending messages to friends For managing friends a member can add a friend from a list of recommended members of our software who are interested in talking about health activities A member may also delete a friend For sending messages a member can send a message to a friend talking about heath topic or whether they want to walk out together Above are the main features for our software More features would be added in as stated in the system requirements 7 2 User Effort Estimation Our system is very easy to use We tried to design it with minimum user effort to accom plish their goal of checking community health activity awareness and entering simple social networking platform The visitor who just wants to check the health activities awareness in certain city and obtain some statistical data 1 NAVIGATION several keystrokes and o
81. rogress Report 15 3 Fut re Work i hg ma Beza sale an aha a ole ah aoa an D Bete AE A 15 4 Breakdown of Responsibility 2044 28 00 8 a be ooh aw Bek ei ened 15 4 1 The breakdown would expected to be as follows 56 56 57 57 59 59 63 64 64 65 66 67 68 68 68 69 69 70 70 71 TI 71 71 71 71 72 72 74 75 82 82 83 85 86 15 4 2 Merging the Contributions from Individual Team Members 16 Reference List 2 Summary of Changes Our project goals from the beginning have changed substantially Below are an itemization of some of the most relevant changes We no longer put an emphasis on integrating user experience and social engineering with the product It was far too ambitious in the planning stages to execute the ideas we set out with given the level of resources within the team User profiles and discussions about users and use cases related to users have been removed There are still references to users and profiles in this report Section 8 preserves a class diagram that could be used with an updated product that has users in it There are deprecated use cases that can be seen in several sections e g 12 1 1 and 12 1 2 These are preserved for posterity There are several removed references to community tracking and providing health suggestions This is far too difficult of a task to carry out We instead narrowed our focus to more simple health tracking of the Un
82. rrectly that message will be sent through the website to the other users inbox and that user will receive a notification that a message was sent Failure x The website is unable to communicate with the server An error message is displayed on the screen and requests that the user check their connection or try again The message is correctly sent but the notification about a new message is not displayed The message appears in the users chat portlet but does not appear in the inbox e UC 9 Changing Account Settings Success The user will opt to change the information needed to log into the website They can change their username or password et cetera If done correctly the database will be updated with the new values Failure x The server is unable to change their username An error message is displayed on the screen and requests that the user try again x The server is unable to change their password An error message is displayed on the screen and requests that the user try again The website is unable to communicate with the server An error message is displayed on the screen and requests that the user check their connection or try again e UC 10 Registering Success x The user will enter their username and password on the registration page and hit submit This will succeed if there is no conflict with the system Failure The website is unable to communicate with the server An error me
83. rsonal Information tab and the system will navigate them to a series of fields where they can change their personal information UC 7 Sending Messages The user wants to message a fellow user or administrator about an issue They will need to interact with the website by creating a new message entering the user information of the person they want to message enter the message they wish to send and then click Send UC 8 Viewing Search History The user wants to view their search history to see things theyve looked at in the past either because they are curious or because they have forgotten something and wish to find it again The user will need to go to their Personal Page via UC 5 and select View Search History tab UC 9 Changing Account Settings The user wants to modify what they see on their personal page or what another user would see on the users page User will head to the Change Account Settings tab and make alterations in various fields corresponding to what needs to be changed UC 10 Registering A user wishes to register a new account so that they can use the program alterna tively an administrator may want to generate additional accounts for whatever reason Account creation requires that the user enter their full name password email birth date and an array of optional information Twitter handle Facebook page et cetera The user will navigate to this screen via either logging out if they are logged in and hitting Sign Up
84. s 2 days 2 days 2 days 16 days 2 days 15 days 14 days 4 days 12 5 days 17 days 6 days 26 days 2 days 28 days 1 day 22 days 21 days 10 days 4 days 4 days 9 days 8 days 3 days Figure 40 Plan of Work 2 Thu 10 31 13 Thu 10 31 13 Fri 11 1 13 Mon 11 4 13 Tue 11 5 13 Wed 10 16 13 Wed 10 16 13 Thu 10 17 13 Thu 10 17 13 Sun 10 20 13 Thu 10 24 13 Mon 11 4 13 Sat 10 26 13 Thu 11 7 13 Mon 11 11 13 Wed 11 6 13 Thu 11 7 13 Thu 11 7 13 Mon 11 18 13 Sat 11 30 13 Fri 11 22 13 Wed 11 27 13 Sat 11 30 13 Thu 12 12 13 Finish November 2013 December 2013 j Gg 19 22 25 28 31 3 6 9 12 15 18 21 24 27 301 3 6 9 12 Wed 11 6 13 yp Fri 11 1 13 Mon 11 4 13 Tue 11 5 13 Wed 11 6 13 Wed 11 6 13 SS Thu 10 17 13 a Wed 11 6 13 Tue 11 5 13 Wed 10 23 13 Mon 1 1 13 Tus 11 26 13 Fri 11 1 13 Thu 12 12 13 Tue 11 12 13 Exi 12 13 13 Eri 12 86 13 Thu 12 5 13 Fri 11 29 13 Wed 12 4 13 Wed 11 27 13 Sun 42 84 13 Tue 12 10 13 Sat 12 14 13 er pa 15 2 Project Coordination and Progress Report e We succeeded in implementing all use cases that weren t related to user specific func tionality e g messaging e The login page works as static content e Website is fully dynamic updates with Twitter data many times an hour e Data analysis is performed through a variety of graphical mechanisms maps charts word cloud et cetera 15 3 F
85. s possible that your results are affected by a small sample size We only have a small number of users collected for a given community so their individual habits can become prominent especially if the time they tweet at during a given day is presumably constant We only collect tweets 3 5 minutes every 15 minutes so if users tweet at the same time every day then we could be biasing our data with that information We have about 300k users from whom we pulled data from over the 1 5 weeks Not all of these users are relevant the problem is many do not have location enabled on their devices and of the ones that do not all of them reside in the United States So as the problem before with gathering many useless tweets we have many tweets from what we can consider to be useless users The interface should be updated with the amount of users who are generating useful tweets as well os we can obtain an idea of how large a subset we are dealing with as far as people who tweet and people who care about their health 51 E ei UC 4 Viewing Map Entering Webpage Initiatiating Page Requesting Twitter ES Requesting Twitter Da 5 Database getTotalTweet Location Categonzation Tweet Content Extraction Prepare marker map controler apdheatmap __ VIEW_MAP A Interact with Display Map Googe Map Figure 20 Use Case 4 Veiw Map Use Case 4 is designed for visualization of heatmap and marker map using Google map service Th
86. s to a constant value and stops or a sort of log t graph again heuristically So the of hashtags is generally greater in our data base than the of tweets themselves and this makes sense 200k hashtags for 10k tweets so about 20 hashtags on average per tweet 59 10 Class Diagram and Interface Specification In this section we describe the framework of our application We originally intended for users to be integrated into the system but after final review it wasn t possible However this design could be implemented later as well with what we have finished to fully integrate users into the system So we elected not to make changes to preserve its functionality 10 1 Class Diagram Controller username String DBConnection password String email String storeNewName String void Checkidentity storeNewPin String void administrator Boolean CreatePersonalProfile storeAge Int wvoid storeSex Bollean void getUsername String UpdatePersonalProfile z storeDOB Int Int Int void setUsername username String void AccessDisplay i storeEmail String void getPsaaword string A MessageManager getLocation String setPassword password string void LocalUser getEmail AceessDisplay TwitterData setEmail getbackPassword username String em STATE_NON_LOGIN int 0 ating strng ni ic sate getbackUsername email String String STATE REGISTERING 2 username String isAdmin
87. ser can press to enter the website 2 Home Screen This is more of the main page of the website The home screen is the central hub for the user to interface with the system and access all of its functionalities Once on the screen the user would be faced with the following components a b Search bar It allows the user to search for cities communities neighborhoods relevant hash tags or other users Main Page This is an area of the screen that displays relevant information based on what is typed into the search bar If there is no text available it will list the following criteria i Available cities This is a list of hot cities relevant to the user ii Facilities This is a list of exercise or health related venues in the surrounding area or in the hot cities iii Search results list This lists the results of a given search when a query is entered into the search bar see 1 Output window This pane of the screen displays information about the community entered into the search bar or selected from the main page This type of information includes the current health of the population in the city notifications about activities in the city and statistical results Account drop down menu This portion of the screen reveals a drop down list of buttons for the user to interact with This screen is mostly about preferences for the users profile i Log in This button brings the user to the login screen where
88. ssage is displayed on the screen and requests that the user check their connection or try again 87 The username is taken by another user The system will request that they try again with another username The password is taken by another user The system will request that they try again with another password e UC 11 Backup Member Data Success The administrator opts to back up the member data and if it is properly copied to another database then the case is successful Failure x The system is unable to communicate with the server An error message is displayed on the screen and requests that the user check their connection or try again The database it is attempting to copy to does not exist or the server is out of space e UC 12 Deleting Member Account Success The administrator opts to delete another members account and it is removed from the system Failure x The system is unable to communicate with the server An error message is displayed on the screen and requests that the user check their connection or try again x The account doesnt exist in the database An error message is sent back to the console e UC 14 Viewing Word Cloud Success The website display all word Cloud integrated and clearly User could change the size with scroll the mouse The word could would move and rotate with the movement of cursor User could get more info with clicking every single word cloud
89. ssible transmission errors UC 8 Viewing Search History This is also inside the personal page Viewing search history can be tested to make sure the most recent member searches can be saved into the database The test should make sure the system handles the search history correctly when there are a lot of entries UC 9 Changing Account Settings This allows the user to change their username or password The test should make sure whether the password can be reset Members can also edit the username The system should be tested on three kinds of changing scenarios changing username changing password and changing both username and password UC 10 Registering Any user can register to be a member to access some additional functions Testing on registration should make sure the user is added into the system member database successfully The registered member should not be registering again which is also a part of the test UC 11 Backup Member Data This test should make sure the member data is backed up periodically to another database in case of runtime failure UC 12 Deleting Member Account As stated in the user case description in previous section deleting member account has two options depending on whether the member wants to permanently delete the account or just make the account invisible to others The test should consider both scenarios UC 13 Viewing Map This is visualization test of the Google MapFragment object The test s
90. sy for the user to find exactly what he or she is looking for Everything they need is clearly labeled and the websites modular design makes it easy for the user to run the services Everything here does not require extensive explanation il Tag Cloud pT ea arte off w es ethest it fitforlife healthy cardio Geng active odin inst health fornfisitieasmotivation TFL ers mie gymlife e weightlgss cycling on GN Zeene Figure 29 Cloud Example 1 78 VAER all Tag Cloud zw CN Ss kee ive healthy trahei nning ogeevsng fitnessn RIA s aut formatigyei RA ec weightloss sm bulk Tit squats swole trainharc sweat instafitness i grind Pusa Rhad fitness REA Sencar 2 i og Mikeplus squagaling Figure 30 Cloud Example 2 allTag Cloud v d COX SS Tit i K 3 instahealth motivation weightloss i swole bulk 1 Figure 31 Cloud Example 3 Results for weightloss p Results gt Pope Top Ail People you fallow a Proks Emily Cooper M D Dt you wonder why diets never work long term read Dr Cooper s new book the Metabolic Storm amazon comidp ROOENN17ML Who to follow Retresh View all Faas ees Comes gt D view summan Reci Te Retweet RFavonte More Ll MES P Guy rien People View all 4 Y Follow Popular account Find hands Trends Change aLiesOdamsTolous Everyday Health Follow Figure 32 Cloud Example 4 Running Word
91. to the database isAdmin This is called by the system in the event that a user attempts to access Admin istrator level privileges Returns the current value of the administrator variable set Admin admin boolean Typical set function It is used when the system needs to change specify Adminis trator privileges for an associated instance of the IdentityList object This would be called when the system or another Administrator is changing user privileges Calling this function updates the value of the administrator variable and forwards the change to the database addUser newUser This function is called when a new user is registering for the system The new User reference is presumed to be an IdentityList object whose set functions were called after the registration button was pushed and populated with values from the form fields The new user is inserted into the database at the end of the function call deleteUser username 58 This function is called when an administrator is attempting to purge a user ac count from the database The function takes a String variable and loops through the database until it finds an IdentityList with a matching username Once found the IdentityList is removed from the database e equals object Object Superclass method of the Java Object class This is overridden and returns TRUE if two variables match each other all classes have this functionality String in cluded
92. tor because they are mostly un ordered the hashtags are always the same for a given tweet set and the dynamic size keeps us from allocating too much or too little for a given search 13 User Interface Design and Implementation The original mockup for our user interface from Report 1 can be seen on the next page Figure 26 Original Mockup 75 Enter your username and password Remember me Forgot Password Register Figure 27 Login UI We kept the same Login UI Search A Home gt Dashboard Total Collected Tweets a 101923 ul Statistics Collected Usable Tweetsa 3319 Total Collected Tweets in US a 1406 76 Estimated number of people in US who exercise a 17575 ww d CX IM Hesith Monitoring Tweet Count for Time of the day MM Northeast EE Midwest Markers IS GOJE a SP o ST d Map data 82013 Google INEGI Obfrenk 4 TermsofUse r Heat Map vse Kaale y s 7 ai Map data 2013 Google INEGI 500kmL __1 TermsofUse Health Monitoring Tweets Distribution of US Regions FOX ul Tag Cloud vee g Ge SS nogss Ze WEN healthy Mspiration an tive cal rdio vali ac grow ining us venta Sen ertoutsoisaneal i ansfomiibiigasmotivation TFLEr squats weightloss wi mie A cycling on ge Zens 77 US South Region B Maine 1 74 GS New Hampshire 2 08 IEI Vermont 0 00 EB Massachusetts 9 38 IEI Rhode Is
93. ust be present in the database Google connection must be open POST CONDITION The information is displayed to the user 1 User goes to the main page to view the data 2 Data if present in the database is pipelined to the website 3 Map fragment is loaded in by Google 4 The map is populated with various tweets after geographic filtering 5 User can view the map breakdown UC 5 Viewing Word Cloud Initiating Actor Member Actor Goal To view results Requirement Addressed REQ1b REQ2a REQ3 REQ8 REQ9 PRE CONDITION Data must be present in the database Twitter website must be available POST CONDITION The information is displayed to the user 22 1 User goes to the main page to view the data 2 Data if present in the database is pipelined to the website 3 Word weight size is calculated using hashtag frequency 4 Weights are updated cloud is shown to the user 6 3 4 Deprecated Use Cases These are use cases that are no longer viable because project goals to implement member functionality could not be completed in the given time and with the team resources UC 5 Accessing Personal Page The user wants to access their personal page in order to peruse what the public can view about them the user User will need to click on their name in the top right of the screen or access the page through a drop down window UC 6 Editing Personal Information The user wants to change his her information User will need to click on the Edit Pe
94. uture Work e Implement the login and user management part e Improve the maintenance routine for our system e Increase functionality such as viewing search history providing individual suggestions 15 4 Breakdown of Responsibility 15 4 1 sending messages adding social networks for members etc The breakdown would expected to be as follows e Twitter API Data Mining Rui Xu Xiaoyu Yu Gradeigh Clark e Google Map sending messages friend invitation 90 Li Xu Yihan Qian Xianyi Gao e Webpage design Data analysis database and server set up main documentation writ ing Gradeigh Clark and Xianyi Gao 15 4 2 Merging the Contributions from Individual Team Members Gradeigh and Xianyi took control of compiling the final report doing the formatting ensur ing consistency and uniform formatting and appearance No extreme issues were encoun tered in compiling the report 16 Reference List 1 Software Engineering Project Health Monitoring Analytics http www ece rutgers edu marsic books SE projects HealthMonitor analytics html 2 Kumar Shamanth Morstatter Fred and Huan Liu Twitter Data Analytics Springer 2013 http tweettracker fulton asu edu tda 3 Kenneth M Anderson Aaron Schram Design and Implementation of a Data Analyt ics Infrastructure in Support of Crisis Information Research NIER Track http epic cs colorado edu wp content uploads icse2011 pdf 4 Twitter Developers
95. ver execute the tweet retrieval script in the back ground Designating a script with nohup when running it in Linux tells the server to continue running the script even if the user who called it is disconnected from it This seemed very ideal for us because that means we could get tweets forever without impunity However two issues arose e The bandwidth limiting from Twitter We can keep the stream open as long as we want but at some point Twitter itself will cut us off anyway We didnt exactly experience this because of the second reason e We were running off of shared hosting from our web host GoDaddy Shared hosting is an inexpensive hosting service where we share the server time and space with other user accounts on the website Running a script with nohup on shared hosting means that youre stealing resources away from everyone else A nohup will run for a little whiel about 7 minutes I believe on GoDaddys shared hosting before they step in and terminate the script forcibly So these two reasons are why we had to resort to cron jobs to pull tweets This data was collected at these discrete time intervals So it s not a continuous open socket of data retrieval the data is retrieved for about 3 5 minutes at a time before the connection is severed So we retrieve tweets for 3 5 minutes every 15 minutes or so So there are gaps where there are no tweets collected at all until the next time marker for another scheduled cron job 48 IS UC
96. y Calculation Word Cloud Configuration Display Word Cloud Figure 11 Use Case Diagram 5 30 7 User Interface Specification 7 1 Preliminary Design This section represents our preliminary design and analysis It is not updated to match the current specs because it was initially planned as a drafting proposal Many things mentioned here are not implemented such as member databasing or logging in Here is the proposed main user interface webpage Q ei wm en Gen y Personal Page 5 y d E e North Main Page 4 Glen Cave ron Si Account p Setting New York City eg Manhattan A Ki Chicogo Output Window Data Analysis Statistical Results Notification Wi Bb E Histogram Chart Others Figure 12 User interface specification draft 1 The visitor user who is not registered can input one of his her interested cities in the search bar and hit search icon The google map will zoom in to the corresponding city and the tweeter users locations will be marked red in the map as shown in Figure 4 1 The tweeter users who mentioned physical activities or concerns about health in their tweets will be marked in different color blue for example The visitor will clearly see how the distribution spreads within this city The map also enables dragging and zooming in out The output window will show the statistical results estimating the percentage of the tweeter users who actually mentioned about health in t
97. y doing the following e data weight ceil log Sfreq Sindex 1 1 2 e Take the logarithm base 10 of the frequency of the hashtag e Add 1 to it to give it some size e Round it up 73 e Add one again e Multiply by 2 This was done experimentally and it seemed to generate satisfactory sizes in relation to the frequency of the hashtags and the size of the tag cloud we used 12 2 Data Structure The system we implemented stores a variety of values that may or may not need to be referenced by the system for any one of the use cases detailed earlier These values go without saying at this point we need to store user data and tweet data and require a data structure that is appropriate to the task Simplistic data structures will not suffice for this project because speed and modularity is a factor in improving the user experience and the system performance The following data structures were under consideration by us e Array This is one of the simplest data structures available We can map an output according to an integer ID But the size of the array is fixed it cant be varied and must be final So if we were to store our data into the array wed need to pre allocate the size and if it got too small wed have to create another array and copy all the data over Not ideal e Queue An ordered list of elements where the principal operations can only be performed on the end tail or front head of the data stru
98. ystem should display a chart showing aggregated tweet data based on the hour of day REQ 7 3 The system should provide statistics based on geographical region REQ 8 3 The system should provide a tag cloud to display the frequency of the used hashtags REQ 9 2 The system should link to Twitter when the tag cloud is clicked REQ 10a 1 The system should have a search function for users REQ 10b 1 The system should allow users to find facilities nearby Table 2 Functional Requirements Table 5 2 Non Functional Requirements Table ID Priority Weight Requirement REQ 11 5 The system should require minimum maintenance at most once per week REQ 12 4 Keep two copies of data for record in case of system failure REQ 13 3 The system should remain functioning in the event of an update to Twitters API REQ 14 2 The software shall present the graph and words in a neat and tidy website Table 3 Non Functional Requirements Table 14 5 3 On Screen Appearance Requirements The following represents the initial draft of what should be major on screen requirements This is meant to be referenced with the figure below The draft and this analysis has since deprecated 1 Welcome Landing Page This is sort of the pre website page It displays information about what the interior of the website contains and gives the user an idea of what he can expect and invites them inside A button would be on the screen somewhere that the u

Health Monitoring Analytics – Final Report

Contents

Download Pdf Manuals

Related Search

Related Contents

Health Monitoring Analytics &ndash; Final Report

Contents

Download Pdf Manuals

Related Search

Related Contents

Health Monitoring Analytics – Final Report