Social Network Analysis: Visualization Tools Dr. oec. Ines Mergel The Program on Networked Governance Kennedy School of Government Harvard University ines_mergel@harvard.edu Content Assembling network data Representing relational data: Sociograms Network visualization as an analysis tool: Interpreting network diagrams Applications Statistical tools Online tools List of online resources Bibliography
Assembling a social network: Data collection Unit of observation: entity on which measurements are taken = actor, from whom we elicit information about ties (actor, pair of actors, relational tie, or event) Modeling unit on a number of different levels: actor, dyad, triad, subset of actor, or network Quantification of the relationship: - directional vs. nondirectional - dichotomous vs. valued Assembling a social network: Data collection Surveys Collect perceptions of interactions List of names or free recall Free vs. fixed choices Ratings vs. complete rankings Observations Face-to-Face interactions: Who talks to whom at a party? Who answers to what kinds of requests on a listserver? Interviews Snowball principle: Who else is important in this network? Face-to-face, or telephone Indirect data Archival records: past political interactions, co-authorship, court records, More reliable: without social ranking Avoid the risk of inaccurate recollections of respondent 2
Representing relational content Network data (Cartwright/Harary 956) Sociometric data: X = population under investigation represented in X x X table Actor-event matrix: X x Y table Arc/node list: ab, ae, ba, bc, bd, ce, dc, ea Board members Y Actor-event matrix Companies X Companies X Network graph: b d a e c Companies X 2 4 4 4 6 2 Sociograms (Moreno, 95) Sociometry : measurement of interpersonal relations in small groups Visual Exploration as an analysis method Representing the formal properties of social configurations Patterns through which perceptions of relationships are structured Useful for: Presenting structures of influence among community elites (Laumann/Pappi, 976) Corporate Interlocks (Levine, 972) Role structures in groups (Breiger et al., 975; Burt 976/982) Interaction patterns in small groups (Freeman et al. 989)
Interpreting a Network Diagram Nodes Relations Lines Arrows Positions Central Peripheral Isolate Groups/Subgroups Sociometric Star Sociometric choices: Whom would you choose as a friend? A B C D E F A -- B -- C -- D -- E -- F -- E F D A B C 4
Example: Connections among Members of Congress websites Visualization of Social Networks Example: UCInet and NetDraw 5
Visualization of Social Networks: Netdraw NetDraw: Sample Davis data. 2-Mode representation 6
Network drawing principles Distance between vertices expresses the strength or number of ties as closely as possible Example: distance between cities on a map = geographic distance Connected nodes are drawn closer together than nodes that are not related. Line length proportional to line values in case of lines with unequal values Don t lump nodes on top of each other or on top of lines they are not connected to Minimize crossing lines Choose automatic drawing mechanisms and then manipulate by hand or redraw several times (de Nooy/Mrvar/Batagelij 25) Example: Network Visualization with Pajek Source: Pajek Homepage (http://research.lumeta.com/ches/map/gallery/index.html) 7
Example: Network Visualization with Pajek Colored by distance from the host Colored by toplevel domain name Colored by ISPs, the city-states of the Internet. Source: Pajek Homepage (http://research.lumeta.com/ches/map/gallery/index.html) Example: Network Visualization with Mage Mage Java Mage Source: http://kinemage.biochem.duke.edu/software/software.html#mage 8
Example: Network Visualization with Visone Analysis and visualization of social networks Example: Network Visualization with Krackplot Circular layout Annealed layout 9
Relational content analysis: Information visualization Source: Prof. Ulrik Brandes, University of Constance (Germany) Touchgraph: Visualization of Interrelated Information http://www.touchgraph.com/
Issue Crawler (govcom.org) Grokker.com See New York Times article: http://www.nytimes.com/25/5/9/technology/9yahoo.html?; http://www.grokker.com/
Online Resources GraphViz (Open Source Graph Drawing Program) ttp://www.graphviz.org/ iknow Inquiring Knowledge Networks on the Web: http://www.spcomm.uiuc.edu/projects/teclab/iknow/ KrackPlot: http://www.andrew.cmu.edu/user/krack/krackplot/krackindex.htmlnetvis Module Dynamic Visualization of Social Networks (MIT): http://www.netvis.org/resources.php Mage - D vector display program which shows "kinemage" graphics (http://kinemage.biochem.duke.edu/kinemage/kinemage.php) Netdraw: http://www.analytictech.com/downloadnd.htm Pajek Program for Large Network Analysis: http://vlado.fmf.unilj.si/pub/networks/pajek/ UCInet: http://www.analytictech.com/default.htm Yahoo usergroup: http://groups.yahoo.com/group/ucinet Visone analysis and visualization of social networks http://www.visone.de Link collection of network visualization tools (especially for Internet data visualization): http://www.caida.org/projects/internetatlas/viz/viztools.html Dr. Ines A. Mergel The Program on Networked Governance Kennedy School of Government Harvard University ines_mergel@harvard.edu 2