Two-mode networks - Affiliations, bibliographic
Contents
1. If you would like you can kaug S CU arate the no de and C dge file S m List T Remove completed automatically Remove all completed gt with I You can later Algorithm Name Date Time Complete Update Network by Merging 05 21 2013 05 30 03 AM Aon n h g h fi v Detect Duplicate Nodes 05 21 2013 05 25 48AM merge t em to et CL or d new v Detect Duplicate Nodes 05 21 2013 05 21 54AM eee network v Network Analysis Toolkit N 05 21 2013 05 14 31 AM OO ewe v Extract Co Author Network 05 21 2013 05 11 37 AM Oe v Load 05 21 2013 05 09 04 AM eee This can make 1t easier to examine and to manipulate or add columns for nodes Constructing a Co Authorship Network Extract all nodes with an attribute strictly above a certain number crm 1010 Data Manager Sr Extract from this number a ISI Data Users amccranie Dropbox INSNA NetworkS ER 473 Unique ISI Records Below a T Extracted Co Authorship Network IN Il Merge Table based on label OW WE WIL remove Numeric Attribute number_of_authored_works SES Text Log Nodes that will be merged Il h h irected ii Text Log Noteworthy nodes that will NOT be n all authors that ui Updated Network nine your data EH Graph and Network Analysis Log EH Merging Report Author information published fewer than 3 papers in this dataset act Nodes Above or Below Value was selected Author s Thomas G Smith2. Documentation http wiki cns iu edu display CISHELL Delet elica TTA oe L s il Balloon Graph prefuse alpha Radial Tree Graph prefuse alpha f eee A 1 mna Drl VxOrd e Radial Tree Graph with Annotation prefuse beta Tree View prefuse beta This algorithm lays out nodes based on the VxOrd force directed layout algorithm Tree Map prefuse beta d Recursive Graph Layout in Force Directed with Annotation prefuse beta Fruchterman Reingold with Annotation prefuse beta 28VxOrd 29 Jj DrL VxOrd e S a Specified prefuse beta New X Position Attribute Name xpos S Circular Hierarchy Remove all completed Gephi New Y Position Attribute Name ypos a Do not cut edges e e Complete 23 31 PM Edge Cutting Strength t e 20 04 PM 57 49 PM Bipartite Network Graph 54 10 AM Cancel 53 48 AM Extracting Word Co Occurrence Network from SDB Data Laying out the network with Drl VxOrd may take some time but once the algorithm 1s complete you will want to keep only the strongest edges so select the Laid out with DrL and run Preprocessing gt Networks gt Extract Top Edges using the shown below File Data Preparation Preprocessing Analysis El Console General L parameters Temporal d Completed cool down Entering crunch stage Finished crunch stage ir PM Entering simmer stage Networks gt Finished simmer s
3. Field edgeid business color directed label labelcolor labelsize labelvisible marriage nodel node2 visible weight width originallabel Value 1 F 176 180 1 false 1 0 0 0 0 255 12 false H n5 nl true 1 0 4 0 Acciaiuoli A Salviati Object all nodes Colour Tornabuoni Guadagni Mridolfi Albizzi Medici o Barbadori e Ginori E Property Operator Value Show Hide Size Show Label Hide Label Format Node Labels Format Edge Labels Node Shape Center Change History Colorize Resize Linear Zoom Level 4 0056 Interpreter e Bischeri eo Lamberteschi Strozzi e Peruzzi amp Castellani Change Label a mw Basic Analysis on the Florentine Dataset OO Graph and Network Analysis Log4603904353876139914 This graph claims to be undirected Add betweenness centrality to the Nodes 16 Isolated nodes 1 dataset Node attributes present label wealth totalities priorates betweenness_centrality Edges 27 Ho self loops were discovered Analysis Networks No parallel edges were discovered i Edge attributes Unweighted amp Undirected gt Nonnumeric attributes Example value Node betweenness aia busines T Did not detect any numeric attributes Then right click to view the new file This network does not seem to be a valued network Remember you will have to Save it if Average degree 3 375 This graph is not weakly connect
4. color 176 180 1 l l ao de To visualize the Mese 12 aa i labelvisible tan LS Florentine network bri as p node2 n8 select de n Q O e o O and NWB LN will launch GUESS zi To change the layout in GUESS select Object all nodes Property yelsize Operator Value you Colour Show Hide Size Show Label Hide Label Change Label can run GEM multiple times to Node Shane Change History Resize Linear Colorize randomly generate a network to your satisfaction as seen to the right Zoom Level 4 0056 eoo Ginori Albizzi File Edit Layout Script View Help Field Value edgeid 1 business F color 176 180 1 directed false Resize the nodes gt e according to family on wealth label n From To Then click The results will look similar to what 1s shown to the right NWB file Users amccranie Desktop sci2 sampledata socialscience florentine nwb GUESS Visualizing the Florentine dataset Object all nodes Colour E Property Operator Value Show Hide Size Show Label Hide Label Change Label Format Node Labels Format Edge Labels Node Shape Center Change History Resize Linear Colorize Zoom Level 4 0056 Interpreter B OO Visualizing the Florentine dataset NWB file Users amccranie Desktop sci2 sampledata socialscience florentine nwb GUESS labelcolor 0 0 0 255 Colorize
5. Dichotomize Multipartite Joining r3 e Merge 2 Networks 01 23 31 PM 01 20 04 PM 00 57 49 PM 11 54 10 AM 11 53 48 AM Extracting Word Co Occurrence Network from SDB Data Once edges have been removed the network top 1000 edges by weight can be visualized by running Visualization gt Networks gt GUESS pim Property lab Format Node Labels Format Edge Labels Node Shape Center Change History Zoom Level 0 30418 Extracting Word Co Occurrence Network from SDB Data In order to make use of the DrL VxOrd force directed layout we applied we need to change to the interpreter at the bottom of the screen and type in the following commands gt gt gt for n in g nodes si n x n xpos 40 PA n y n ypos 40 gt gt gt resizelinear references 2 40 gt gt gt colorize references gray black gt gt gt resizeLinear weight 1 2 gt gt gt g edges color 127 193 65 255 gt gt gt Interpreter Graph Modifier Extracting Word Co Occurrence Network from SDB Data Note GUESS will not necessarily display the graph 1n the middle of the screen you may have to scroll around the screen to find the graph 128 128 128 255 false 2 0 Necessit 0 0 0 255 12 false n667 Necessit 1 black 2 true 2 0 3040 7920000000004 76 0198 5584 679999999999 139 617 gt gt gt fr m
6. den Dunn han Hermpakob Henning Kien Julie Thompson 2 e fogram Associate Michael J e e santi N Goldston David ham Feeney SR pall Char Christopher B Konum Kai Q 9 Ple E From Sci User Manual Figure 5 1 Co authorship network of Katy B rner tJ Pacheco Roberto Wales Jimy Cockentt Matthew Mons Barend Kirkendall Nancy J 95 J David Study Director Thomas J Taylor Martha M Edge Size amp Color Number of Time Co Authored 14 4 1 E 14 1 uud Roes Peter Jan a Chicester Christine Barris Nicolas GUESS supports the repositioning of selected nodes Multiple nodes can be selected by holding down the Shift key and dragging a box around specific nodes The final network can be saved via GUESS File gt Export Image and opened in a graphic design program to add a title and legend The image above was created using Adobe Photoshop Node clusters were highlighted and increased in size the label font size was increased for emphasis and a legend was added to clarify the significance of node and edge size and color http sc12 wiki cns 1u edu dis lay SCIZTUTORIAL S5 1 Individual Level Studies Micro emporal Visualization SGER Collabora Research Mapping the Structure and Evolution of Sustainability Soence Reeeech s TLS Towards a Macroscope for Science Policy Decision Making NSF Workshop on Knowledge Management and Visusitzzton Tools in Support of Discovery nr
7. 4 CSV file C Users dapolley AppData Local Temp temp Preprocessed eet End Deeded Nebeod Ei with normalized article title stati Extract Bipartite Network Extract Paper Citation Network affili Extract Author Paper Network abst New Extract Co Occurrence Network T Extract Word Co Occurrence Network pubi artic Extract Co Author Network med Extract Reference Co Occurrence Bibliographic Coupling Network issn own Extract Document Co Citation Network ml j 3 Detect Duplicate Nodes T Els Update Network by Merging Nodes CD Given a table this algorithm creates a co word occurrence Database network where the strength of edges between papers represents the similarity of their abstracts Node Identifier Column L Tex Delimiter Aggregate Function File C Users dapolley Desktop sci2 Browse Jarticle title Algorithm Name Date Time Complete Lowercase Tokenize St 11 26 2012 10 39 21 AM Extracting Word Co Occurrence Network from SDB Data To see more information about your network run Analysis gt Networks gt Network Analysis Toolkit File Data Preparation Preprocessing Analysis Modeling Visualization R Help El Console Temporal Oo T Data Manager Geospatial i 4 E CSV file C Users dapolley AppData Local Temp temp Preprocessed Topical ES 4 EF with normalized article title ml journal info medline ta false Networks Network Analysis Toolkit NAT ta Co Word
8. A Research Field 1990 2012RecoveryCorrectedcitesandnames deleted 1283466bd59201e0e9105946a2fa90d7 isi 29g TH d Last Saved 3 4 13 12 43 26 AM pe 4 Al File Path v Dropbox dropbox cache 2013 05 18 1990 201 s deleted 1283466bd59201e0e9105941 pm 2 1990 2012RecoveryCorrectedcitesandnames deleted 1283466bd59201e0e9105946a2fa90d7 isi gt C F Rus J mE Dennis M Y THERAPEUTIC OPTIMISM oe E tify j Professionally Defined Funk R R li d 168 47 0 TI Implications of multiple co ocurring psychiatric problems for longterm Lt Mi f P Pd te a PERSONAL JOURNEY Patient Defined 138 39 e recovery managemen 2 SO ALCOHOLISM CLINICAL AND EXPERIMENTAL RESEARCH LA English stract CT 27th Annual Meeting of the Research Society on Alcoholism 004 Uses refers to both 5 36 1096 E s a P n n s 2 e Refers to neither B a L A A 12 399 CO ism 9 C1 Ctr Addict amp Mental Hlth Toronto ON Canada 3 PU LIPPINCOTT WILLIAMS amp WILKINS 4 PI PHILADELPHIA 5 PA 530 WALNUT ST PHILADELPHIA PA 19106 3621 USA 6 SN 0145 6008 7 J9 ALCOHOL CLIN EXP RES 8 JI Alcoholism NY 323588 11 none Unicode UTF 8 Uns LF 20 401 939 350 456 amp PSYCHOLOGISTS rre IN PSYCHIATRY DEPARTMENTS PSYCHIATRISTS From McCranie 2013 unpublished dissertation in progress Image in middle above 1s largest single component of the paper citation network of recovery research literature fr
9. E 9 Us Object nodes based on gt E Property number of authored works Operator gt Value Colour Show Hide Size Show Label Hide Label Change Label Format Node Labels Format Edge Labels Node Shape Center Change History Resize Linear Colorize Zoom Level 0 58372 USING A MAP OF SCIENCE TO LOCATE RESEARCH The Map of Science is a visual representation of 554 sub disciplines within 13 disciplines of science and their relationships to one another shown as points and lines connecting those points respectively Over top this visualization 1s drawn the result of mapping a dataset s Journals to the underlying sub discipline s those journals contain Mapped sub disciplines are shown with size relative to the number matching journals and color from the discipline For more information on maps of science see http mapofscience com As of the Sci2 v1 0 alpha release there 1s a plugin for Sc12 that allows users to visualize their own data overlaid on the Map of Science Load the FourNetSciResearchers 1si file in the ISI flat format Fille Data Preparation Preprocessing Analysis Modeling Visualization R Help El Console D IHI Data Manager help with your analyses have questions about datasets or would like to suggest enhancements and new features Primary investigators are Katy B rner Indiana University and Kevin W Boyack SciTech Strategies Inc The Sci2 tool wa
10. Save to EndNOl Web Save to EndNot RefMan ProCite O All records on pu CJ plus Abstract Y Save to other Reference Software C Save 7 Records 1 to 500 i Full Record i Save to BibTeX i M pius Cited Reference i Save to HTML Save to Plain Text Save to Tab delimited Win 586 records matched your query of the 14 495 418 in the data limits you selected Save to Tab delimited Mac m Network Science Journal Editors e Alessandro Vespignani e Lada Adamic e Nosh Contractor e Stanley Wasserman e Thomas Valente e Garry Robins e Sanjeev Goyal e Ulrik Brandes You will see some errors These are items that are not journal articles but books and chapters You could restart your search and exclude them Sci2 Tool Peyton 2003 EC COMPLEX EVOLVING VIII CAMBRIDGE Article Book Chapter 10 1017 CBO9780511614385 005 sgoyalQyessex ac uk 167 Thomson Reuters Web of Knowledge 978 0 521 84271 6 BDM98 GROUP FORMATION IN ECONOMICS NETWORKS CLUBS AND ICOALITIONS English LOCAL INTERACTION MODEL SOCIAL INTERACTIONS STATISTICAL MECHANICS CORRELATED EQUILIBRIA EVOLUTIONARY GAMES ECONOMIC NETWORKS LONG RUN COORDINATION DIFFUSION NEIGHBORS 46 B 2005 CAMBRIDGE UNIV PRESS THE PITT BUILDING TRUMPINGTON ST CAMBRIDGE CB2 1RP CAMBS ENGLAND Goyal S reprint author Univ Essex Dept Econ Colchester CO4 3SQ Essex England Univ Essex Dept Econ Colchest
11. and then sums the result For binary data this ends up with a measure of concurrent attendance membership For valued data you can use a minimums method taking the smallest value of a pair of actors ties to events and summing that with all other actors Other places to go Knoke and Yang Social Network Analysis 2008 Faust Using Correspondence Analysis for Joint Displays of Affiliation Networks in Carrington Scott and Wasserman Roberts J M Correspondence analysis of two mode network data Social Networks 22 65 72 2000 Finding Social Groups A Meta analysis of the Southern Women Data In Ronald Breiger Kathleen Carley and Philippa Pattison eds Dynamic Social Network Modeling and Analysis Washington D C The National Academies Press 2003 Borgatti S P and M G Everett Network analysis of 2 mode data Social Networks 19 3 243 269 1997 Recent extensions to 2 mode data e Blockmodeling P Doreian V Batagelj Generalized blockmodeling of two mode network data Social Networks 200 e Large two mode networks Matthieu Latapy Clemence Magnien Nathalie Del Vecchio Basic notions for the analysis of large two mode networks Social Networks Volume 30 Issue 1 January 2008 Pages 31 48 ERGM p models P Wang K Sharpe G Robins and P Pattison Exponential random graph models for affiliation networks Social Networks 31 1 2009 pp 12 25 A demonstration and remind
12. deciding on the population boundaries of each Is Important Duality of Actors and Events FLORA DOROTHY OLIVIA lE PEARL RE N E11 RUTH E9 e MYRNA N EVELYNZ donee NJ CHARLOTTE ELEANOR E CTI gl COMO EVELYN LAURA THERESA BRENDA EHARLOTTE FRANCES ELEANOR PEARL ISUTH VERNE MYRNA KATHERINE SYLVIA NORA HELEN DOROTHY OLIVIA FLORA El E2 E3 E4 ES E6 E7 E8 E9 El E11 E12 E13 E14 LA A LA LA Hp l LL d 1 1 1 1 1 1 1 1 1 1 1 Ll 1 1 1 1 T b 35 3b Lil LA LA LA LA LA LA Pajek shadow 0 00 1 00 EVELYN LAURA THERESA BRENDA CHARLOTTE FRANCES ELEANOR PEARL RUTH VERNE MYRNA KATHERINE SYLVIA NORA HELEN DOROTHY OLIVIA FLORA X EVELYN THERESA BRENDA CHARLOTTE FRANCES ELEANOR PEARL RUTH artite Matrix VERNE MYRNA KATHERINE SYLVIA NORA H H 4 HH T E HELEN DOROTHY OLIVIA FLORA Two Mode Data Basic Measures Properties you should know Rates of participation W amp F 312 313 a size of events W amp F313 314 UCINET now has some nice measures you can easily compute 2 Mode Cohesion Measures for davis dataset 1 Z 3 4 5 6 7 Density Avg Dist Radius Diameter Fragmenta Transitiv Norm Dist Matrix 1 0 352 NES LETS 9 000 4 000 0 000 Dol 0 647 NOTE If
13. 2 Jan 29 2013 When Temporal Data Week 4 Feb 12 2013 What Topical Data Theory Mid Term to be taken by Feb 18 2013 at 5pm EST Welcome by K Welcome by Katy B rner 39 VIDEO Welcome Week 5 Feb 19 2013 With Whom Trees Exemplary Visualizations 9 46 VIDEO SLIDES Week 6 Feb 26 2013 With Whom Networks Overview and Terminology 16 38 VIDEO SLIDES Workflow Design 19 41 m VIDEO SLIDES WU LI A Burst Detection 14 14 i VIDEO SLIDES r d E Self Assessment LINK applied to answer when questions Week 7 Mar 5 2013 Dynamic Visualizations amp Deployment va Y EL Hands on Introduction by Ted Polley 42 VIDEO Next Tassnacsal Mae P e kaa AICP Te ed ae Monfilaa f3 A4 AY Micro Individual Scientist e Node Size amp Color lotal T ores Herr ll Bruce W Hu Wetria Bonnie emele oss unns 1 6 menr emi pas 22 KeMWeima 3 176 1 Moore Chris Chen Chaomei e e Carsten Ebad Ursyn Anna e Izquierdo Ebroul Roberts Jonathan C 2 9 rina Gordon e g Zhang Jian J e 0 o Muhammad ea e amp com bre amp e i e 9 E Weeder Marc Melton Wiliam Packer Abel B Baroch Amos B Menssen oe Moeller Erk Lewis Suzanna Soe R Astbumar Michael Van Muligen Erik Mons Albert van Ommen
14. and 24 hours of NIH tutorials in this tool Gallery Pm Fear Epidemics Tool EpiC The EpiC Tool supports the custom analysis modeling and visualization of data streams such as diffusion patterns of the H1N1 virus over geographic space Gallery Types and level of Analysis Micro Individual Meso Local Macro Global dine IUI 5 600 records 10 000 lt records Individual person and Larger labs centers K All of NSF all of USA their expertise profiles universities research Ml of science domains or states Temporal haf ing topi 113 Years of physics When i in 20 year Research Career trajectory of one Mapping a states PNAS publications individual intellectual landscape Geospatial Analysis Where Knowledge flows in VxOrd Topic maps of Chemistry research NIH funding Topical Analysis Base knowledge from What which one grant draws ggigi iggflyNvvrheoeoBr rr U O UE u _ _ _rrrueumtEEpueceuu lt mn lt ERGEIE r s Xerwork Analysis NSF Co PI network of Co author network NSF s core competency be ion LIU enner p I Why or why not to use Sc12 Pros Friendly interface otrong extensive documentation with sample workflows Powerful parsing capabilities for both general and specific file types Works well in using and producing common data products so movement between programs is not difficu
15. fragmentation is gt 0 the graph is disconnected All measures based on lengths of geodesics are computed within components Density is the number of ties divided by n m where these no of rows and cols in matrix Avg Dist is the average geodesic path length in the bipartite graph within components Radius is the smallest eccentricity in the bipartite graph within components Diameter is the length of the longest geodesic in the bipartite graph within components Transitivity is the no of quadruples with 4 legs divided by no with 3 or more legs in bipartite graph Norm Dist is Avg Dist divided into minimum possible in bipartite graph of given node set sizes Two Mode Data Network Density For the actors density is the average number of events attended by all pairs of actors g g A DA iT CREE V 2 i lt j You can do something similar for event density Two Mode Data Network Centrality Actor Degree Centrality in an affiliation network is the total number of actors contacts that the actor has through its attendance at all events oum that actor s row in the coattendence matrix You can also do something similar for events g Grad x ij l 1 Centrality for Davis 2 Mode Centrality Measures for ROWS of davis 1 2 3 4 Degree Closeness Betweenne Eigenvect tota M These are routines available in wr M UCINET under Network gt 2 Qo umo me oye vom ram Mode Networks and prov
16. maximum data values are given in the legend Color Discipline See end of PDF for color legend Now consider this with the NetworkScienceEditors txt file See the spread of the area coverage That s what they were going for Note where the largest circle is reflects the large number of contributions of Vespigani Topical Visualization Generated from 473 Unique ISI Records 2 193 out of 227 publications were mapped to 222 subdisciplines and 13 disciplines May 21 2013 06 01 AM CEST Chemistry Ne E Social Scie pg ON Ka Medical Specialties m ica gt T Ea 2aCiBarncas 2008 The Regents of the University of Califomia and SciTech Strategies Map updated by SciTech Strategies OST and CNS in 2011 Legend Area How To Read This Map Circle area Fractional Joumal Count 29 09 The UCSD map of science depicts a network of 554 subdiscipline nodes that Unclassified 34 are aggregated to 13 main disciplines of science Each discipline has a distinct Minimum 0 color and is labeled Overlaid are circles each representing all records per Maximum 60 unique subdiscipline Circle area is proportional to the number of fractionally Color Discipline 28 assigned records Minimum and maximum data values are given in the legend See end of PDF for color legend USING THE SCHOLARLY DATABASE ON A TOPIC AREA Word Co Occurrence Network from the abstracts of articles from MEDLINE with the keyw
17. nodes according to how M many seats in government the 51 259 family holds a Barbadori Field Value color 0 51 51 255 fixed false height 6 0827625 image label Barbadori File Edit Layout Script View Help gt gt gt for nin g nodes n strokecolor n color gt gt gt Graph Modifier a Mm Visualizing the Florentine dataset 8090 NWB file Users amccranie Desktop sci2 sampledata socialscience florentine nwb GUESS NE File Edit Layout Script View Help gannas di 051 51 255 v Q Switch to the Interpreter at the bottom of the GUESS non window vate oe 9 i 0515155 Q Type the following ME true 55 v 6 0827625 115 0 77 5 commands for n in g nodes n strokecolor n color n strokecolor n color Note after typing the first line hit the Jab key and after the second line hit Enter and the commands Will be executed Learn more about GUESS script options by looking at the sample 0 files in the sampledata folders and by visiting http graphexploration cond org documentation html Add the family labels to the nodes select Object Then Click The family names will then appear next to their corresponding nodes B O O Ginori Albizzi NWB file Users amccranie Desktop sci2 sampledata socialscience florentine nwb GUESS File Edit Layout Script View Help Visualizing the Florentine datase
18. valued Average degree 147 644 EI Scheduler Remove From List Remove completed autor Algorithm Name Date W Network Analysis Toolki 11 26 2 Extract Word Co Occurr 11 26 20 W Lowercase Tokenize St 11 26 2012 Extract Top Nodes Extract Nodes Above or Below Value Delete Isolates Extract Top Edges Extract Edges Above or Below Value Remove Self Loops Trim by Degree MST Pathfinder Network Scaling Fast Pathfinder Network Scaling Snowball Sampling n nodes Node Sampling Edge Sampling Dichotomize Multipartite Joining Merge 2 Networks 15348AM asm C WW Data Manager 4 CSV file C Users dapolley AppData Local Temp temp Preprocessed a EE with normalized article title _ 4 A Co Word Occurrence network i Graph and Network Analysis Log Extracting Word Co Occurrence Network from SDB Data Apply Visualization gt Networks gt DrL VxOrd and words that are similar will be plotted relatively close to each other Set the parameters to those shown below File Data Preparation Preprocessing Analysis Modeling Visualization R Help 2 Console General ka Manager EL Temporal P csv file C Users dapolley AppData Local Temp temp Preprocessed Geospatial gt EF with normalized article title Delete Isolates was selected Networks GUESS _Implementer s Patrick Phillips l Integrator s Patrick Phillips Topical b Cytoscape
19. Endicott J 1976 ARCH GEN PSYCHIAT V33 P766 Ridgway P 2001 PSYCHIATR REHABIL J V24 P335 Kessler RC 1994 ARCH GEN PSYCHIAT V51 P8 Citation Count 406 257 177 164 159 155 144 LS 113 111 99 98 95 93 91 39 88 88 87 85 83 80 79 74 In ISI Dataset book yes too early too early too early too early book yes yes yes too early yes yes yes too early too early yes too early yes too early yes too early yes does not fit search The Demonstration e Basic Features of Sci2 e Citation and co authorship networks with special emphasis on preparing and cleaning data e Topical analysis and visualization s Available data and ways to pull it into Sci2 Questions amp discussion THE BASIC FUNCTIONALITY The Basics lt Opening Data Data reparation e Preprocessing e Analysis Modeling e Visualization p LLL M4 M 0 IoeioA Ss gt N gt RARA AEAEAEEeA gt AoA gt I IITTT EEEMLEEEEZEO OOO ng zo _ m Gm File Data Preparation Preprocessingeit liiciioiodeling Visualization R Help a National Science Foundation under Grant No SBE 0738111 and IS 0715303 and the Jame S McDonnell Foundation See the Science of Science homepage http sci2 wiki cns iu edu for documentation and screenshots Please visit https sci2 cns iu edu user ask php if you need help with your analyses have questions about datasets or would li
20. Implementer s Thomas G Smith Integrator s Thomas G Smith Documentation http wiki cns iu edu display CISHELL Extract Nodes Above or Below Value EJ Scheduler 1 M Algorithm Name Date Time Complete Extract Nodes Above or Bel 05 21 2013 05 34 11 AM Network Analysis Toolkit N 05 21 2013 05 32 21 AM Update Network by Merginc 05 21 2013 05 30 03 AM Detect Duplicate Nodes 05 21 2013 05 25 48 AM Detect Duplicate Nodes 05 21 2013 05 21 54 AM Network Analysis Toolkit N 05 21 2013 05 14 41 AM n o m YARRA f 4 4 4 Enter 2 instead of the 3 you see Run the NAT again and see what has changed Let s visualize in GUESS File Edit Layout Script View Help Valente Tw Mon Field Value J _edgeid 266 color 176 180 directed false label labelcolor 0 0 0 255 labelsize 12 labelvisible false node 1 n4 TL node2 n56 number 0o 1 originallabel visible true QU weight 1 0 width 4 0 Constructing a Co Authorship E Network 800 Nodes above 2 000000 by number_of_authored_works GUESS t p L e P ee P T e Y a e iatronero L 1 a R 6A elemy M x a D t 9 i Barrat A ei paat ras R 4 e o 17 Qui e gt Qon Pe e Adsmic ta D j 02 Si Contrector Ns IE dh i L N e U t u 0 wagner d a a j
21. Occurrence network Extract Word Co Occurrence Network wi Implementer s Micah Linnemeier Integrator s Micah Linnemeier Unweighted amp Undirected Documentation http wiki cns iu edu display CISHELL Extract Word Co Occut Weighted amp Undirected Unweighted amp Directed Input Parameters Node Identifier Column article title Weighted amp Directed Text Delimiter article title false abstract true Class edu iu nwb analysis extractdirectednetfromtable extractdirectednetwork ExtractDirectedNetworkAI gorithmFactory Tr E Scheduler mL Remove From List Remove completed automatically Remove all completed Algorithm Name Date Time Complete Extract Word Co Occurr 11 26 2012 11 54 10AM EE Lowercase Tokenize St 11 26 2012 11 348 AM E Extracting Word Co Occurrence Network from SDB Data Look at the resulting network with the Network Analysis Toolkit Delete the 1solate nodes by running Preprocessing gt Networks gt Delete Isolates File Data Preparation Preprocessing Analysis Modeling Visualization R Help El Console General K Nodes 955 Temporal g Isolated nodes 579 Node attributes presenti Geospatial K Topical k Edges 70500 Networks No self loops were discoverer No parallel edges were discovered Edge attributes Did not detect any nonnumeric attributes Numeric attributes min max mean weight 1 4 1 21599 This network seems to be
22. Time Complete Extract Co Author Network 05 21 2013 05 11 37AM eee Load 05 21 2013 05 09 04 AM 900 Graph and Network Analysis Log8847484992994669308 This graph claims to be undirected Nodes 1385 Isolated nodes 2 Node attributes present label number of authored works times cited Edges 115252 Mo self loops were discovered Edge attributes Did not detect any nonnumeric attributes Numeric attributes min max mean number 1 34 1 46848 weight 1 34 1 46848 This network seems to be valued Average degree 176 6314 This graph is not weakly connected There are 6 weakly connected components 2 isolates The largest connected component consists of 1288 nodes Did not calculate strong connectedness because this graph was not directed Density disregarding weights 8 1355 Additional Densities by Numeric Attribute 1083 Constructing a Co Authorship Network Sort by number of articles and by label Look for Wasserman for instance Notice that he is in there more than once with different name spellings Adamic 1s too This is a problem So choose the network again in the data manager and Abernethy J Adamic L Adamic La Adar E Adeolu A Adevemi dor LO GO SJ Od M 8 OC Detect Duplicate Nodes Detects duplicate nodes by comparing their attributes Attribute to compare on label x L Merge when thi
23. Unit Affiliations Data and Bibliographic Research ICPSR University of Michigan Ann Arbor oummer 2015 Instructor Ann McCranie Affiliations A Relational in three ways Wasserman and Faust pg 295 1 Show how actors and events are related to one another 2 Events create ties among actors 3 Actors create ties among events Note You are no longer considering pairs of actors but instead are considering subsets of actors Classic Study Southern Ladies Davis Gardner Gardner aaa aa MS 1941 Deep South a Social sees Anthropological Study of ae risk 1000 3 O0 111111110000 0 Caste and Class i 4 123211115202 Se 5 CHARLOTTE 0011101000000 0 6 FRANCES 0010110100000 0 7 ELEANOR 0000111100000 0 See also Freeman Linton 8 PEARL 00000101100000 A 9 RUTH OU O D 6 LL D Lb lI 1 DO 9 O 9 2003 Finding Social 10 VERNE 00000011100100 11 MYRNA 0 000000111010 0 Groups A Meta Analysis of 12 KATHERINE 0000000111011 1 1 3 SYLVIA O D V 0 D OQ LDL bL LL Li 9 qd 1 1 the Southern Women Data 1 00 sora 00000110111111 T9 HELEN O 0 0 0 0 Q 1 LO L 1 Lk 0 Q 16 DOROTHY O 000000110000 0 17 OLIVIA O 0 D OG D DU 0 L 0 r0 O D 18 FLORA 0000000010100 0 Two Mode Data Issues Think carefully about the assumptions behind the collection of actors and events Does belonging to the same organization count as an affiliation What about belonging to the same type of organization How about sharing the same attitude Also
24. cience Scheduler F Remove completed automatici Subtitle Generated from 361 Unique ISI Records Rrrr T Journal Column Journal Title Full ka Map of Science via Jour 11 26 2012 Load 11 26 2012 Scaling Factor auto Simplified Layout Show Export Window El Console General Y ta Manager Documentation http wiki cns iu edu display CISHELL Data4 Temporal 1 ISI Data C Users dapolley Desktop sci2 sampledata scientometrics i Found old style ISI Web Of Knowledge file Geospatial B 361 Unique ISI Records Found old style ISI Web Of Knowledge file T gt Found old style ISI Web Of Knowledge file Found old style ISI Web Of Knowledge file Topical 4 Map of Science via Journals The original 361 records have been processed to remove duplicz ui TCU ha sli Balloon Graph prefuse alpha in Gi Snc ai Esse CK Cr Topical Visualization Generated from 361 Unique ISI Records 80 out of 112 records were mapped to 182 subdisciplines and 13 disciplines November 26 2012 04 37 PM EST Health Professionals Math amp 2 Q lt ge o 3 v 9 z d m Ei E i cv i o o 2 a IC DS rale 3 E Biotechnolagy o dod Sy E E oe R o Electrical Engineering Brain Research E Computer Science Chemical Weaning Cio eo Infechaus Di ta 1 2 ys bet eb Acer Social Scityfes o Medical Specialties a agro o PCS Biology g Earth Sciences Humanities 2008 The Regents oft
25. e Sci2 tool was developed by Chin Hua Kong Joseph Biberstine Thomas G Smith David M il StopwordList C Users dapolley Desktop sci2 configuration stopwords txt Coe Micah W Linnemeier Patrick A Phillips Chintan Tank and Russell J Duhon It uses the Cyberinfrastructure Shell http cishell org developed at the Cyberinfrastructure for Network Science Center http cns iu edu at Indiana University Many algorithm plugins were derived from the Network Workbench Tool http nwb cns iu edu New Separator n owner Please cite as follows Sci2 Team 2009 Science of Science Sci2 Tool Indiana University and SciTech Strategies http sci2 cns iu edu Loaded C Users dapolley Desktop MEDLINE_master_table csv S status pubmodel E Scheduler mL Remove From List Remove completed automatically Remove all completed volume issue T issn Algorithm Name Date Time Complete published_month article title medline pagination 0 IEEE affiliation 1 Extracting Word Co Occurrence Network from SDB Data Run Extract Word Co Occurrence Network and set the parameters as shown below File Data Preparation Preprocessing Analysis Modeling Visualization R Help G d Remove ISI Duplicate Records D nt mi j Remove Rows with Multstudinous Fields
26. earch field e Can be used to identify conflict and consensus in science e Can be used to trace the development of Ideas m Why might you want to do this 1f you aren t into bibliometrics e Get a handle on an intellectual field he people in that field s The academic work they produce and what Is most cited he clusters and divisions of the people and ideas in the field e Maybe you just want to find the biggest stars and greatest hits there are now a lot of tools to help you do that Studying Bibliographic Networks These are special types of social networks e Instead of two scholars being linked together by their relationship such as being friends or naming each other as colleagues two scholars are linked in a bibliographic network by the work they produce together s n a citation network two papers are linked by the works they both cite or the works that both cite them These are a type of affiliation network technically a type of two mode network Mode Network Actors Events SCI TOOL DEMONSTRATION Download the tool and documentation https sc12 cns 1u edu amccrani indiana edu Logout Sci Tool A Tool for Science of Science Research amp Practice Home Download The Science of Science Sci Tool is a modular toolset specifically designed for the study of science It supports the temporal geospatial topical and network a
27. ed you want to use it after you have There are 2 weakly connected components 1 isolates closed the program Did not calculate strong connectedness because this graph was not directed The largest connected component consists of 15 nodes Density disregarding weights 8 225 Now use the NAT to get basic stats on the network Analysis gt Networks gt Network Analysis Toolkit It will be reported in your console window and as a output log 1n the EN data manager CO AUTHORSHIP OF A NEW JOURNAL m Constructing a Co Authorship m Network e Acquire the data We ll use ISI but you have many choices Load and prepare the data e CLEAN and verify the data e Duplicates loops errors misspellings etc e Analysis e Visualization Constructing a Co Authorship Network These data were gathered by searching ISI Web of Science a well established bibliographic database that has wide coverage in science social science and the humanities Details of each individual search are noted below Please note that you might get slightly different results when you attempt to search particularly if ISI Web of Science has added journals to their database in the meantime e Once you have conducted your search you can export the search results into a number of formats To create the ISI files like we we use in this this tutorial you will need to save them as Plain Text Note that you can only exp
28. ended them Correspondence analysis uses singular value decomposition of a normalized version of your g x h matrix PEARL RUTH Bl E7 EB D G DE DOROTHY VERNE 0 5 E11 SYLVA B 4 THERINE 1 5 2 5 Factions 1n 2 mode data Starting fitness 0 000 Final fitness 0 490 Correlation to ideal 0 490 Blocked Adjacency Matrix 1 2 3 4 5 6 7 8 9 10 11 12 13 14 El E2 E3 E4 ES E6 E7 E8 E9 E10 E11 E12 E13 E14 1 EVELYN 1 000 1 000 1 000 1 000 1 000 1 000 1 000 1 000 2 LAURA 1 000 1 000 1 000 1 000 1 000 1 000 1 000 3 THERESA 1 000 1 000 1 000 1 000 1 000 1 000 1 000 1 000 4 BRENDA 1 000 1 000 1 000 1 000 1 000 1 000 1 000 5 CHARLOTTE 1 000 1 000 1 000 1 000 6 FRANCES 1 000 1 000 1 000 1 000 7 ELEANOR 1 000 1 000 1 000 1 000 8 PEARL 1 000 1 000 1 000 9 RUTH 1 000 1 000 1 000 1 000 10 VERNE 1 000 1 000 1 000 1 000 11 MYRNA 1 000 1 000 1 000 1 000 12 KATHERINE 1 000 1 000 1 000 1 000 1 000 1 000 13 SYLVIA 1 000 1 000 1 000 1 000 1 000 1 000 1 000 14 NORA 1 000 1 000 1 000 1 000 1 000 1 000 1 000 1 000 15 HELEN 1 000 1 000 1 000 1 000 1 000 16 DOROTHY 1 000 1 000 17 OLIVIA 1 000 1 000 18 FLORA 1 000 1 000 Density matrix 1 0 625 0 074 2 Os e 0 537 Turning Two Mode into One Mode oymmetric Valued Data Most common way to convert Is called cross products his method takes each entry of the row for actor A and multiplies it times the same entry for actor B
29. er CO4 350 Essex England Economics Business amp Economics 11 Learning in Networks 11 WOS 000313909500005 1 0 Demange G Wooders M 10 1017 CBO9780511614385 Users amccranie Dropbox INSNA NetworkScienceEditors txt has no jounal column J9 The Cite Me As field may be invalid he original 473 records have been processed to remove duplicate unique ISI IDs leaving 473 records rote log to var folders tr qtdkk6b17f5 ltd7700vjppgr0000gn T isiduplicateremoverlog5835257979007378045 txt Loaded Users amccranie Dropbox INSNA NetworkScienceEditors txt El Scheduler Remove From List Remove completed automatically Remove all completed Algorithm Name Date Time Complete v Load 05 21 2013 05 09 04 AM eee Constructing a Co Authorship Network dio Data Manager Tom M ISI Data Users amccranie Dropbox INSNA NetworkSc 473 Unique ISI Records Constructing a Co Authorship Network eoo Temporal DI z Geospatial it Data Manager _ _rrr__ 3 Topical U 9 000313909500005 1 0 Demange G Wooders M 10 10 Networks gt Network Analysis Toolkit NAT inie Dropbox INSNA NetworkSd amccranie Dropbox INSNA NetworkScienceEditors txt ha _ n rds As field may be invalid E 3 he original 473 records have been processed to remove duplicate unique ISI IDs leaving Unweighted amp Und
30. er e Sci2 There data available for you to download on the course webpage B AN INTRODUCTION TO BIBLIOGRAPHIC NETWORK ANALYSIS Scientometrics 1s an independent and diverse field Berg 9 Literally measuring and analyzing Scientometrics science all kinds tini e Today were just in a tiny little corner of bibliometric analysis Co authorship e Citation e But this field covers all sorts of ways of quantifying and evaluating scientific efforts Bibliographic Network Analysis Is often focused on citation or authorship patterns that can be found within fields Co authorship is pretty clear shared authorship of an article Citation analysis can be thought of differently Two common ways e bibliometric coupling where two citing articles are similar to the extent they cite the same literature and co citation analysis where cited articles are similar to the extent they are cited by the same citing articles Hummon and Doriean Social Networks Volume 11 Issue 1 March 1989 Pages 39 63 Citation networks are an interesting case because at their most basic they are directed acyclic data A paper cites a paper that was written earlier that older paper can never cite anything published later Bibliographic network analysis e Can detect subcommunities in research fields e Can and has been be used to map science e Can be used to identify prominent actors in a res
31. es Primary investigators are Katy B rner Indiana University and Kevin W Boyack SciTech Strategies Inc The Sci2 tool was developed by Chin Hua Kong Joseph Biberstin O florentine9071067285680920294 nwb Smith David M Coe Micah W Linnemeier Patrick A Phillips Chintan Tank and Nodes Duhon It uses the Cyberinfrastructure Shell http cishell org developed att l f TET Cyberinfrastructure for Network Science Center http cns iu edu at Indiana id int label string wealth int totalities int priorates int Many algorithm plugins were derived from the Network Workbench Tool http 4 Acciaiuoli 18 2 53 nwb cns iu edu 2 Albizzi 36 3 65 3 Barbadori 55 14 8 Please cite as follows 4 Bischeri 44 9 12 FI E gt OP ND Sci2 Team 2009 Science of Science Sci2 Tool Indiana University and SciTe 5 Castellani 28 18 22 http sci2 cns iu edu 6 Ginori 32 9 8 Mee 7 Guadagni 8 14 21 Load was selected 8 Lamberteschi 42 14 e the Florentine Documentation http wiki cns iu edu display CISHELL Data Formats 9 Medici 103 54 53 Loaded Users amccranie Desktop sci2 sampledata socialscience florentini 18 Pazzi 487 0 i 11 Peruzzi 49 32 42 View with was selected 42 Pucci 3 1 0 L datas et im 14 Salviati 10 5 35 Input Parameters 13 Ridolfi 27 4 38 15 Strozzi 146 29 74 T Ru 16 Tornabuoni 48 7 8 L e E schedule pin o view the file right
32. he University of California and SciTech Strategies Map updated by SciTech Strategies OST and CNS in 2011 Legend Area How To Read This Map Circle area Fractional record count Rr The UCSD map of science depicts a network of 554 subdiscipline nodes that Unclassified 22 are aggregated to 13 main disciplines of science Each discipline has a distinct Minimum 0 color and is labeled Overlaid are circles each representing all records per Maximum 98 unique subdiscipline Circle area is proportional to the number of fractionally Scaling factor 0 5076673 0 85 assigned records Minimum and maximum data values are given in the legend Color Discipline See end of PDF for color legend The journals titles are used to determine which records fit into what sub discipline You can view the journal titles found and those not found from the data manager of Sc12 A single Journal can belong to more than one sub discipline and thus so can the record associated with that journal So the circle sizes are proportional to the number of fractionally assigned records iology 92008 The Regents of the University of California and SciTech Strategies e i St ies OST d CNS in 2011 are aggregated to 13 main disciplines of science Each discipline has a distinct color and is labeled Overlaid are circles each representing all records per unique subdiscipline Circle area is proportional to the number of fractionally assigned records Minimum and
33. ide mene 0429 QI 0 047 0 22 appropriate normalizations for M A m S i fe the values If you just forced 17 cuum S 0 585 0 008 0 07 your data into a bipartite graph it would not be normalized correctly for the number of Degree Closeness Betweenne Eigenvect actors and events 2 Mode Centrality Measures for COLUMNS of davis 1 El 0 167 0 524 0 002 0 142 2 E2 0 167 0 524 0 002 0 150 3 E3 0 333 0 564 0 018 0 253 4 E4 0 222 0 537 0 008 0 176 5 BS 0 444 0 595 0 038 0 322 6 E6 0 444 0 688 0 065 0 328 7 E7 0 556 0 723 0 130 0 384 8 E8 0 778 0 846 0 244 0 507 9 E9 0 667 0 786 0 226 0 2579 10 E10 0 278 0 550 0 011 0 170 11 E11 0 222 0 537 0 020 0 090 12 E12 0 333 0 564 0 018 0 203 13 E13 0 167 0 524 0 002 0 113 14 E14 0 167 0 524 0 002 0 113 Closeness and Betweeness We can also get closeness and betweenness measures for two mode networks but we have to change the way they are normalized in order to reflect the fact that we have nodes that by definition can not be adjacent to one another Read Extending Centrality by Everett and Borgatti in Carrington for more but think of this Actor closeness centrality hn 2g 2 Ce 1 Correspondence Analysis See Faust Chapter in Carrington for excellent description Correspondence Analysis looks at the correlations between two sets of variables and is used to locate actors and events simultaneously actors near events they attended and events near actors who att
34. ies L M Inc The Sci2 tool was developed by Chin Hua Kong Joseph Biberstine Thomas G Smith David Gallery M Coe Micah W Linnemeier Patrick A Phillips Chintan Tank and Russell J Duhon It uses the umm ent Cyberinfrastructure Shell http cishell org developed at the Cyberinfrastructure for Network E CT ims du 5 Science Center http cns iu edu at Indiana University Many algorithm plugins were derived gt T s k from the Network Workbench Tool http nwb cns iu edu z K Ke 2D n AME Please cite as follows 1 M Watch the movie about ClShell Powered B Scheduler B tools on the SciVee Making Science Visible S website by clicking on the image above Remove completed automatically Remove all completed Algorithm Name Date Time Complete Science of Science Tool Sci n i ET The Sci Tool was specifically developed for science policy makers and researchers that study science by V Google Citation UserID 05 18 2013 025137PM mn scientific means It supports the temporal geospatial topical and network analysis and visualization of A Load 05 18 2013 025125PM m scholarly datasets at the micro individual meso local and macro global levels There exists a 112 MI Load 05 18 2013 024833 PM M page user manual a continuously updated Sci Tool wiki
35. im gnodes g nodes kaz n x n xpos 40 ia n y n ypos 40 gt gt gt resizelinear references 2 40 gt gt gt colorize references gray black gt gt gt resizeLinear weight 1 2 gt gt gt g edges color 127 193 65 255 gt gt gt g edges color blue gt gt gt rere 127 193 65 255 gt gt gt _ j ijijb bu i xz R ab G_ _gG_ _GgG m gt gt onde Sl ard gt Sci2 Manual gt Science of Science Sci2 Tool Manual Search duction ing Started yrithms Tools and Plugins kflow Design 1016 Workflows 1016 Science Studies amp Online Services nding the Sci2 Tool avant Datasets and Tools rences i2 FAQ 1dix 1 Introduction to Networks 1dix 2 Glossary 1dix 3 ClShell Algorithms dix 4 Date Format dix 5 Sci2 Release Notes 9 Science of Science Sci2 Tool Manual 6 Added by Micah Linnemeier last edited by Ted Polley on Dec 05 2012 view change Science of Science Sci2 Tool Manual v1 0 alpha Download an offline copy of Sci2 v1 0 alpha manual 44 MB Updated July 16 2012 A tool for science of Science research amp practice A communi fons Watch the movie about ClShell Powered tools on the SciVee Making Science Visible website by clicking on the image above http www scivee tv node 27704 Just the beginning check the Sci2 website for more
36. irected orship Network records Weighted amp Undirected b rote log to var folders tr qtdkk6b17f5 1td7700vjppgr0000gn T Unweighted amp Directed isiduplicateremoverlog5835257979007378045 txt Weighted amp Directed Loaded Users amccranie Dropbox INSNA NetworkScienceEditors txt Extract Co Author Network was selected Implementer s Timothy Kelley Integrator s Timothy Kelley Documentation http wiki cns iu edu display CISHELL Extract Co Author Network 28Text Files 29 Input Parameters File Format isi Now look at the network with the NAT E scheduler Note the number of Remove From Ust Remove completed automatically Remove all completed nodes H Algorithm Name Date Time 96 Complete 117 900 jGraph and Network Analysis Log8847484992994669308 Extract Co Author Network 05 21 2013 05 11 37 AM This graph claims to be undirected Load 05 21 2013 05 09 04 AM ___ r_ Nodes 1385 Isolated nodes 2 Mode attributes present label number of authored works times cited Edges 115252 Mo self loops were discovered Mo parallel edges were discovered Edge attributes Did not detect any nonnumeric attributes Numeric attributes min max mean number 1 34 1 46848 weight 1 34 1 46848 This network seems to be valued Average degree 176 6314 This graph is not weakly connected There are 6 weakly connected co
37. ke to suggest enhancements and new features Primary investigators are Katy Borner Indiana University and Kevin W Boyack SciTech Strategies Inc The Sci2 tool was developed by Chin Hua Kong Joseph Biberstine Thomas G Smith David M Coe Micah W Linnemeier Patrick A Phillips Chintan Tank and Russell J Duhon It uses the Cyberinfrastructure Shell http cishell org developed at the Cyberinfrastructure for Network Science Center http cns iu edu at Indiana University Many algorithm plugins were derived from the Network Workbench Tool http nwb cns iu edu y Remove all completed Algorithm Name Time Load 05 18 2013 06 05 43 PM M Load 05 18 2013 06 05 33 PM W Load 05 18 2013 06 05 23 PM ISI Data C Users amccrani Downloads S 787 Unique ISI Records Visualizing the Florentine dataset Data Acquisition amp Preprocessing e Examine the data s Load the data into NWB e Data Analysis Modeling amp Layout e Since the data is formatted as a network with various attributes added to the nodes and edges only simple analysis will be conducted on this network e Data communication amp Visualization Layers Visualize the Florentine network with GUESS Visualizing the Florentine dataset Florentine families related through business ties Specifically recorded financial ties such as loans credits and joint partnerships and marriage alliances Node att
38. lt R Pajek visone Gephi Can handle very large datasets Can be customized with plug ins Cons Fewer ready implemented algorithms than statnet UCINET or Pajek Less flexible visualization within the program Adding plug ins can be daunting For more information about other software that could be of interest for the study of science see section 8 2 of the Sci user manual for listing and discussion Data suited for this tool Specific file formats e Publications e Network Formats e Refer BiblX enw e GraphML xml or graphml BibTeX e XGMML xml e SI Web of Science e Pajek NET net e Scopus e NWB nwb e Google Scholar e Scientometric Formats e Google Citation e ISI isi Funding Bibtex bib NSF Award Search e Endnote Export Format enw e NIH RePORTER e Scopus csv scopus e Scholarly Database e NSF csv nsf e Other Formats e Pajek Matrix mat e reeML xml Edgelist edge CSV csv e Databases e Scholarly Database http sdb cns iu edu search Free Online Course http ivmooc appspot com cours Schedule Pre Questionnaire Week 1 Jan 22 2013 Visualization Framework amp Workflow Information Visualization MOOC INDIANA UNIVERSITY E Week 2 Jan 29 2013 When Temporal Data Home Schedule Announcements My Profile Forum FAQ Week 3 Feb 5 2013 Where Geospatial Data Course Week2 Lesson 1 Week
39. mponents 2 isolates The Largest connected component consists of 1288 nodes Did not calculate strong connectedness because this graph was not directed Density disregarding weights 8 1355 Additional Densities by Numeric Attribute Constructing a Co Authorship Network Geospatial E Tal i Data Manager zm x Topical Dy 0003 13909500005 1 0 Demange G Wooders M 10 10 Networks K Network Analysis Toolkit NAT inie Dropbox INSNA NetworkS amccranie Dropbox INSNA NetworkScienceEditors txt ha rds As field may be invalid i he original 473 records have been processed to remove duplicate unique ISI IDs leaving Unweighted amp Undirected K orship Network records Weighted amp Undirected gt t rote log to var folders tr qtdkk6b17f51td7700vjppgr0000gn T Unweighted amp Directed isiduplicateremoverlog5835257979007378045 txt Weighted amp Directed Loaded Users amccranie Dropbox INSNA NetworkScienceEditors txt Extract Co Author Network was selected Implementer s Timothy Kelley Integrator s Timothy Kelley Documentation http wiki cns iu edu display CISHELL Extract Co Author Network 3 2 8Text Files9629 Input Parameters File Format isi Now look at the network with the NAT El Scheduler Note the numb CT of Remove From List Remove completed automatically Remove all completed Algorithm Name Date
40. nalysis and visualization of scholarly datasets at the micro individual meso local and macro global levels Registration required Download lt la kete News 2012 e Jul 24 The latest Sci Tutorial for the National institutes of Health Portfolio Analysis Symposium is now available Jun 13 The Sci Science of Science Tool v1 0 alpha release is now available supports Mac OS X Cocoa frameworks platform and featuring new visualizations Google Scholar Citation reader R Documentation Ask An Expert Testimonials Developers A Q 9 e am owe ar P r i P 9 e i tee Cm EN oe t oo ker Hry i A OPE LR EC x WARUM e N tmm EN E F K E aer a od Qa 4 2 Rat tirar 7 M ms A K ei rates ec IL GLA UGENI data E d oro I e m Mapping Topics and Topic Bursts in PNAS Start Stop 4 Have a question Ask an Expert A Science of science tool Cyberinfrastructure Shell CIShell X CIShell supports the plug and play of datasets and algorithms and their bundling into custom tools that serve the specific needs of a user group or research community M ig MD ue et 4 It has been applied to develop diverse custom tools see ClShellasaplatform E mu below Feel free to take plugins from any of these tools to for your tool design your personal dream tool Provided by the Cyberinfrastructure for Network Science Center a
41. om 1991 2012 Top right image is Pathfinder Network Scaling of the most cited works in transformed co citation network color coded by content analysis of articles Bottom right 1s co authorship network X coded by disciplinary field Data prep and analysis in Other Sci2 visualization in visone with added legends and y sychiatrists graphic elements in Inkscape kd CA e North American Psychiatrists simple but important Most highly cited works American Psychiatric Assoc 1994 DIAGN STAT MAN MENT Anthony WA 1993 PSYCHOSOC REHABIL V16 P11 Hamilton M 1960 J NEUROL NEUROSUR PS V23 P56 American Psychiatric Assoc 1987 DIAGN STAT MAN MENT Kay SR 1987 SCHIZOPHRENIA BULL V13 P261 Deegan PE 1988 PSYCHOSOC REHABIL V11 P11 American Psychiatric Assoc 2000 DIAGN STAT MAN MENT Davidson L 1992 BRIT J MED PSYCHOL V65 P131 Jacobson N 2001 PSYCHIATR SERV V52 P482 Frank E 1991 ARCH GEN PSYCHIAT V48 P851 Overall JE 1962 PSYCHOL REP V10 P799 Andreasen NC 2005 AM J PSYCHIAT V162 P441 Mueser KT 2002 PSYCHIATR SERV V53 P1272 Liberman RP 2002 INT REV PSYCHIATR V14 P256 Folstein MF 1975 J PSYCHIAT RES V12 P189 Keller MB 1987 ARCH GEN PSYCHIAT V44 P540 Young SL 1999 PSYCHIATR REHABIL J V22 P219 Beck AT 1961 ARCH GEN PSYCHIAT V4 P561 Harrison G 2001 BRIT J PSYCHIAT V178 P506 Harding CM 1987 AM J PSYCHIAT V144 P727 Deegan P 1996 PSYCHIATR REHABIL J V19 P91
42. ord mesothelioma in the title If you have registered for the Scholarly Database then go to http sdb cns iu edu and login SCHOLARLY DATABASE Cyberinfrastructure for Network Science Center SLIS Indiana University Bloomington Non IU User IU Users mustlogin using the Central Email Authentication Service CAS the standard IU authentication system Please click the button below to proceed to the IU login page Go to IU Login Password Login If you do not have an account type in nwb indiana edu and nwb for the password Do a keyword search in Title for mesothelioma and check MEDLINE Search o Edit Profile o About o Logout o Search Creators Title mesothelioma Abstract AH Text First Year 1865 v Last Year 2012 v C Clinical Trials 1900 2012 MEDLINE 1865 2010 NEH 1970 2012 NIH 1972 2012 NSF 1952 2010 USPTO 1976 2010 If multiple terms are entered in a field they are automatically combined using OR So breast cancer matches any record with breast or cancer in that field You can put AND between terms to combine with AND Thus breast AND cancer would only match records that contain both terms Double quotation can be used to match compound terms e g breast cancer retrieves records with the phrase breast cancer and not records where breast and cancer are both present but the exac
43. ort 500 records at a time If you have a search with more records save them 500 records at a time and splice the files together removing the notations for beginning and ending files from the second and third etc set of records you add to your first file Results 586 Show 10 per pag rs 4 i Page 1 of 59 G gt Pi Output Records Step 1 Step 2 Step 3 How do export to bibliographic management software O Selected Records on page i O Authors Title Source i Print E mail Add to Marked List Save to EndNot Web Save to EndNote Ref Man ProCite O A f C plus Abstract Taie MORA P i Save to other Reference Software Save Records 4 to 500 i Full Record Save to BibTeX M pius Cited Reference Save to HTML Save to Plain Text Save to Tab delimited Win 586 records matched your query of the 14 495 418 in the data limits you selected Save to Tab delimited Mac WEB OF KNOWLEDGE DISCOVERY STARTS HERE 3 THOMSON REUTERS Go to mobile site Signin Marked List 0 My EndNote Web My ResearcherlD My Citation Alerts My Saved Searches LogOut Help All Databases Select a Database Web of Science Additional Resources vers tg Search Author Search Cited Reference Search Advanced Search Search History ooi B MORE d F OR NEW USERS Web of Science now with books Get EndNote X6 Now Search Store your references and PDFs and find full text in
44. out datasets or would like to suggest enhancements and new features Primary investigators are Katy Borner Indiana University and Kevin W Boyack SciTech Strategies Inc The Sci2 tool was deweloned hu Chin Hua Kono lnsenh Biherstine Thama mith David h Coe Micah W Linne Cyberinfrastructure Sh Science Center http from the Network Worl tam The file C Users dapolley Desktop MEDLINE_master_table csv can be loaded using one or more of the following formats WO Please select the format you would like to try Please cite as follows Load as 5c Tem 209 Scie I http sci2 cns iu edu Standard csv format NSF csv format Scopus csv format E Scheduler Algorithm Name Complete xtracting Word Co Occurrence etwork from SDB Data Normalize the titles by running Preprocessing gt Topical gt Lowercase Tokenize Stem and Stopword Text and select Abstract File Data Preparation Preprocessing Analysis Modeling Visualization R Help El Console General O Ww Data Manager Foundation See the Scit i http sci2 wiki cns iu edu for Ei CSV file CA Users dg documentation and scri Geospatial E sci2 cns iu edu user ask php if you need help with your analyses Topical gt Lowercase Tokenize Stem and Stopword Text and new features Networks E Primary investigators are Katy BOrrrer Indiana University and Kevin W Boyack SciTech Strategies z i Inc Th
45. r __T zni Creatwe Metaphors to Strmuzste New Approsches to Visusimng Understanding and Rettiniong Large Repostones of S MH 12 Il Visual ang Network Dynamics Competition at fe ntemational Conference on Network Scenos 2007 m F ro mm S C I U S e r Mapping Soence Extitat st the 233d Noral Meeting amp Exposrtion of the Amencan Chemical Sooety in Chicago IL ESSI n E n scr sso e Manual Figure 5 3 m n n May 2008 te mstional Wierkahc and Conference on Network See RO Horizontal Bar Graph of KatyBorner nsf Net Aoreench A Large Soale Network Analysis Modeling and Visustzzion SC Workshop Te Role of Social Network Research in Enabling Cyber Mapping Chemestry 2004 2005 2006 2007 2008 200 Legend Area MEM 52 423 How To Read This Map Pred ENABLE Leamung th CAREER Visuakzing Know 2003 9 2010 2011 Area size Awarded Amount to Date nta 17 474 This temporal bar graph visualization represents each record as a horizontal inimum 11 890 5 825 bar with a specific start and end date and text label on its left side The area aximum 1 120 926 0 of each bar encodes a numerical attribute value e g total amount of funding ext label Title k so Year s Bars may be colored to present categorical attribute values of records i http sc12 wiki cns 1u edu display SCI2TUTORIAL 5 1 Individual Level Studies Micro Micro Trends for Individua Scientist mre Alessandro Vespignani Sternbe
46. rg Distinguished Professor Northeaster computational sciences Network Science Epi Physics Verified email at neu edu Homepage Export articles File format BibTeX 2002 2004 2006 2008 2010 Legend How To Read This Map BibTeX Area size Weight This temporal bar graph visualization represents each record as a horizontal j E dN P err E a bar with a specific start and end date and a text label on its left side The area n t by of each bar encodes a numerical attribute value e g total amount of funding L Export the ore low E Export ali articles Alessandro Ves Easier ena rd E 08 Yea ris Bars may be colored to present categorical attribute values of records RefMan See end of PDF for color legend Title Auth Epidemic spreading in scale free networks R Pastor Satorras A Vespignani 2001 Physical review letters 86 14 3200 3203 The architecture of complex weighted networks A Barrat M Barthelemy R Pastor Satorras A Vespignani 2004 Proceedings of the National Academy of Sciences of the United States of From Sci User Manual http sci2 wiki cns 1u edu displav SCI2TUTORIAL 5 2 Institution Level Studies icro Four Scientists Joint Co Authorship Network Eugene Garfield Node Size amp Color Numl f Papers DEI 127 35 5 1 127 Stanley Wasserman Alessandro Was pignani I Edge Size amp Color Nu Co Autho 33 9 1 1 33 Meso
47. ributes e Wealth Each family s net wealth in 1427 in thousands of lira e Priorates The number of seats on the civic council held between 1282 1344 e otalities Number of business marriage ties in complete dataset of 116 families Edge attributes Marriage T F amp Business T F oubstantively the data include families who were locked in a struggle for political control of the city of Florence around 1430 Two factions were dominant in this struggle one revolved around the infamous Medicis the other around the powerful Strozzis More info is at http svitsrv25 epfl ch R doc library ergm html florentine htm l and Padgett amp Ansell 1993 http nome uchicago edu jpadgett papers published robust pdf Visualizing the Florentine dataset E Console B tit Data Manager Com Welcome to the Science of Science Tool Sci2 N NWE file Users amccranie Desktop sci2 sampled The development of this tool is supported in part by the Cyberinfrastructure for Network Science center and the School of Library and Information Science at Indiana University the National Science Foundation under Grant No SBE 0738111 and IIS 0715303 and the James S McDonnell Foundation See the Science of Science homepage http sci2 wiki cns iu edu for documentation and screenshots Please visit https sci2 cns iu edu user ask php if you need help with your analyses have questions about datasets or would like to suggest enhancements and new featur
48. s properties mergelsiAuthors properties Browse a start In actual analysis you etes Krrrrrrrr r Update Network b would want to be as precise peer as possible Documentation hl Sa OK But for how CNTRL Click om List Remove completed automatically Remove all completed gt the Network and the Merge w Table and go to DATA rrr Merginc e rrr SA PREPARATION gt UPDATE Z Detect Duplicate Nodes 05 21 2013 05 25 48 AM v Detect Duplicate Nodes 05 21 2013 05 21 54AM NETWORK BY MERGING v Network Analysis Toolkit N05 21 2013 05 14 41 AM wi Extract Co Author Network 05 21 2013 05 11 37 AM IN 0 D DN v Load 05 21 2013 05 09 04 AM For this you need the aggregation function file shown Constructing a Co Authorship Network EI Console Ca biol Data Manager Com I ISI Data Usersfamccranie Dropbox INSNA NetworkSc Update Network by Merging Nodes vr ag mr mentation http wiki cns iu edu displav CISHELL Detect Duplicate Nodes t Duplicate N or s Micah L Update a network by merging nodes ementer s Mi rator s Mic mentation hi Parameters e when this si bute to com ber of shared Aggregation Function File Users amccranie Desktop sci2 sampledata scientometrics properties mergelsiAuthors properties Browse e notice whet It should be lower than 1t was te Network Pmenter s rator s Wei mentation
49. s developed by Chin Hua Kong Joseph Biberstine Thomas G Smith David M Coe Micah W Linnemeier Patrick A Phillips Chintan Tank and Russell J Duhon It uses the Cyberinfrastructure Shell http cishell org developed at the Cyberinfrastructure for Network Science Center http cns iu edu at Indiana University Many algorithm plugins were derived from Ld Pleas The file C Users dapolley Desktop sci2 sampledata scientometrics isi FourNetSciResearchers isi can be loaded using one or more of the following formats Please select the format you would like to try EIL Load as ISI flat format ISI scholarly format Date Time C Load 11 26 2012 04 24 51PM Algorithm Name Complete To visualize dataset overlaid on the Map of Science run Visualization gt Topical gt Map of Science via Journals _ 2 Sci2 Tool H File Data Preparation Preprocessing Analysis Modeling Visualization R Help l Wrote log to Ci Users dapolley AppData Local Temp isiduplicateremoverlog4672608363523779963 txt Loaded C Users dapolley Desktop sci2 sampledata scientometrics isi FourNetSciResearchers isi Map of Science via Journals was selected Implementer s David M Coe Integrator s David M Coe Documentation http wiki cns iu edu display CISHEL Locate the journals from a table on the UCSD Map of S
50. s similar Mad Create notice when this similar 9 85 Number of shared first letters 2 K ior s Mic meie 6 Implementer s Micah Linnemeier Integrator s Micah Linnemeier a Documentation http wiki cns iu edu display CISHELL Detect Duplicate Nodes 8 El Scheduler 18 Algorithm Name Date 19 Detect Duplicate Nodes 05 21 2013 20 V Network Analysis Toolkit N 05 21 2013 Extract Co Author Network 05 21 2013 23 v Load 05 21 2013 1 2 9 829 2 236 1 272 1 272 Time Complete 05 21 4 AM D 05 14 41 AM rr __ 6 05 09 04 AM via Data Manager Lo n 398 473 605 860 862 ISI Data Users amccranie Dropbox INSNA NetworkS c a 473 Unique ISI Records T Extracted Co Authorship Network EB Graph and Network Analysis Log Sn Author information Constructing a Co Authorship E Network 0 Data Manager Com Documentation htto wiki cns iu edu displav CISHELL Detect Duplicate Nodes Et ISI Data Users amccranie Dropbox INSNA NetworkSc O Update Network by Merging Nodes Detect Dun Author s Micah L Update a network by merging nodes Implementer s Mi Integrator s Mica Documentation hl Now examine the merge table Look at the two reports MM 3 arameters It s clearly not perfect but it s fifa Number of shared Aggregation Function File Users amccranie Desktop sci2 sampledata scientometric
51. seconds with EndNote in Topic X6 and EndNoteSync Try it now Example oil spill mediterranean mu AND 4 in Author 4 Select from Inde Training and Support Example O Brian C OR OBrian C m e Download Quick Find usn Need help finding papers by an author Use Author Search Recorded Training wg Fi AND in Publication Name Select from Inde Example Cancer OR Journal of Cancer Research and S Clinical Oncology What s new in Web of Knowledge Add Another Field gt gt e Data Citation Index Discover use and cite research data More information e More of What s New Featured Tips Limits To save these permanently sign in or register e Visualize citation connections at a Ti glance with Citation Mapping Timespan view demo l All years updated 2013 05 17 e Identify citation trends graphically N z with Citation Report view demo From 1900 to 2013 default is all years e Howto update your Researcher Records processed from to ID profile a Citation Databases Customize Your Experience V Science Citation Index Expanded SCI EXPANDED 1955 present a aS Pe eee ee ee ee ee at eee ee Results 586 Show 10 per page 3 La i Page 1 of59 GoD ip bi Output Records Step 1 i Step2 i Step3 Howdolexport to bibliographic management software Q Selected Records on page O Authors Title Source Print Add to Marked List
52. source int target int marriage string busi 9 we al 9 1 d ie p s 2 ST ph TT click on the network II 9 2 T p 5 3 TA ai a 6 3 p TS e Algorithm Name Date Time 193 T T th d T View with 05 21 2013 02 07 43 AM 113 F T 1n C ra a Manager 9 7 4 Ta WT y v v Load 05 21 2013 02 06 26 AM B4 F T eT and select view as B5 F T p H a taba E G55 A7 F TE Select a text editor 8 7 T T 16 7 T F 118 pa ap 10 9 F T gt 13 9 T F 149 SD 16 9 T T 14 1B T F 15 11 T F 15 13 T F Nodes id int label string wealth int totalities int priorates int Acciaiuoli 10 2 53 2 Albizzi 36 3 65 3 Barbadori 55 14 0 4 Bischeri 44 9 12 5 Castellani 20 18 22 6 Ginori 32 9 0 7 Guadagni 8 14 21 8 Lamberteschi 42 14 0 9 Medici 103 54 53 10 Pazzi 4870 11 Peruzzi 49 32 42 12 Pucci 310 13 Ridolfi 27 4 38 14 Salviati 10 5 35 15 Strozzi 146 29 74 16 Tornabuoni 48 7 0 UndirectedEdges source int target int marriage string business string 9 1 T E 62 T F FTT p 9 2 T F E spen Visualizing the Florentine dataset eoo NWB file Users amccranie Desktop sci2 sampledata socialscience florentine nwb GUESS File Edit Layout Script View Help Tornabuoni Medici Field Value edgeid 22 Q lo business T D a
53. t Indiana University G Sci2 Tool E ERA File Data Preparation Preprocessing Analysis Modeling Visualization R Help Learn more about existing CIShell powered tools below Console att Data Manager Ca Welcome to the Science of Science Tool Sci2 The development of this tool is supported in part by the Cyberinfrastructure for Network Science center and the School of Library and Information Science at Indiana University the National Science Foundation under Grant No SBE 0738111 and IIS 0715303 and the James S McDonnell Foundation See the Science of Science homepage http sci2 wiki cns iu edu for gt A CSV file C Users amccrani AppData Local Temp temp Pi E Google Citation User Id E EF Citation Table OTJZPjJOAAAAJ 1 FA Citation Table 25_HO_wAAAAJ Network Workbench Tool NWB The NWB Tool supports researchers educators and practitioners interested in the study of biomedical social documentation and screenshots Please visit https sci2 cns iu edu user ask php if you need Citation Table USCXAPsAAAAJ and behavioral science physics and other networks It help with your analyses have questions about datasets or would like to suggest enhancements Ss Citation Table RINNTAMAAAAJ i 5 2 and new features ES Citation Table 13K1 czwAAAAJ comes with a 77 page user manual n Primary investigators are Katy Borner Indiana University and Kevin W Boyack SciTech Strateg
54. t phrase The importance of a particular term in a query can be increased by putting a and a number after the term For instance breast cancer 10 would increase the importance of matching the term cancer by ten compared to matching the term breast Extracting Word Co Occurrence Network from SDB Data We will only download the first 1000 results to minimize the runtime for the algorithms used in this workflow Make sure to check MEDLINE master table since that will have all of the bibliographic data we need for this analysis Search Edit Profile About Logout Download Results ay Download all Data Dictionary T Sample File Download 1000 records starting atrecord 1 from the following databases DI MEDLINE Database MEDLINE author table T MEDLINE co author table nwb format ES MEDLINE master table MEDLINE MeSH heading table E MEDLINE MeSH qualifier table CEED Your download limit will initially capped at 2000 records at a time To increase this limit please email cns sdb dev l uiulist indiana edu Extracting Word Co Occurrence Network from SDB Data save the file somewhere on your computer for use later in this tutorial 1n this case 1f you can t sign on to SDB then you will find a copy in the DATA folder You have chosen to open 1 sdb download zip which is a Compressed zipped Folder from http sdb cns iu edu What should Firefox do with this file Open with Windo
55. tage ins Ser pns totar energy Layout calculation completed in 22 seconds not it Writing out solution to inFile icoord Total Energy 67 8669 Program terminated successfully Geospatial ETETETT Extract Top Edges was selected Author s Thomas G Smith Implementer s Thomas G Smith Integrator s Thomas G Smith Documentation http wiki cns iu edu display Cl El Scheduler Remove From List Remove completed autor Date 11 26 2 Algorithm Name i Extract Top Edges 4 DrL VxOrd i DrL VxOrd N Delete Isolates Network Analysis Toolki Extract Word Co Occurr Lowercase Tokenize St 11 26 2012 11 26 2012 11 26 2012 11 26 2012 11 26 2012 gy 4 30416e 007 IO ADA DOT 11 26 2027 U12511PN 0 Modeling Visualization R Help ois Data Manager 4 CSV file C Users dapolley AppData Local Temp temp Preprocessed 4 with normalized article_title 4 S Co Word Occurrence network EB Graph and Network Analysis Log 4 With isolates removed S Laid out with DrL Extract Top Nodes Extract Nodes Above or Below Value Delete Isolates Extract Top Edges Extract Edges Above or Below Value Remove Self Loops Trim by Degree MST Pathfinder Network Scalin Extract the top N edges from a graph based on a given attribute Fast Pathfinder Network Scalinc Snowball Sampling n nodes Extract N top edges 1000 Node Sampling Edge Sampling Extract bottom edges instead
56. ws Explorer default v Do this automatically for files like this from now on N m Extracting Word Co Occurrence Network from SDB Data The topic similarity of basic and aggregate units of science can be calculated via an analysis of the co occurrence of words in associated texts Units that share more words in common are assumed to have higher topical overlap and are connected via linkages and or placed in closer proximity Extract Word Co Occurrence Network creates a weighted network where each node Is a word and edges connect words to each other where the strength of an edge represents how often two words occur in the same body of text together This algorithm is a shortcut for extracting a directed network using Extract Directed Network and then performing bibliographic coupling using Extract Reference Co Occurrence Bibliographic Coupling Network Extracting Word Co Occurrence Network from SDB Data Open Sci2 and load the MEDLINE master table csv file as a Standard CSV file File Data Preparation Preprocessing Analysis Modeling Visualization R Help El Console O iit Data Manager Science Foundation under Grant No SBE 0738111 and IS 0715303 and the James S McDonnell Foundation See the Science of Science homepage http sci2 wiki cns iu edu for documentation and screenshots Please visit https sci2 cns iu edu user ask php if you need help with your analyses have questions ab
Download Pdf Manuals
Related Search