Visualizing Social Networks Ryan Yee 1 Plan Introduction

  • Slides: 25
Download presentation
Visualizing Social Networks Ryan Yee 1

Visualizing Social Networks Ryan Yee 1

Plan Introduction and terminology Vizster Node. Trix Mat. Link Applications to Multi-agent systems 2

Plan Introduction and terminology Vizster Node. Trix Mat. Link Applications to Multi-agent systems 2

Terminology Graph Model G = (V, E) Graph Visualization Network / Node-Link Adjacency Matrix

Terminology Graph Model G = (V, E) Graph Visualization Network / Node-Link Adjacency Matrix 3

Node-Link Vertices represented by nodes Edges represented by links 4

Node-Link Vertices represented by nodes Edges represented by links 4

Adjacency Matrix Vertices represented by rows/columns Edges represented by cells Edges are directed 5

Adjacency Matrix Vertices represented by rows/columns Edges represented by cells Edges are directed 5

Node-link vs. Adjacency Matrix Much easier to trace path in node-link Much easier to

Node-link vs. Adjacency Matrix Much easier to trace path in node-link Much easier to compare degree in matrix 6

Problems…? Social data is: Locally dense Traditional node-link excellent for overall structure but less

Problems…? Social data is: Locally dense Traditional node-link excellent for overall structure but less so for dense Adjacency excellent for dense structures but not effective for overall structure 7

Vizster: visualizing Online Social Networks (Heer, Boyd) Visualization of friends Spring-based layout algorithm Focuses

Vizster: visualizing Online Social Networks (Heer, Boyd) Visualization of friends Spring-based layout algorithm Focuses on subset of known friends Spring strength based on vertex degree Designed for finding clusters 8

Vizster: 1 & 2 Degree 9

Vizster: 1 & 2 Degree 9

Vizster: path finding 10

Vizster: path finding 10

Vizster: Focus 11

Vizster: Focus 11

Vizster: Searching 12

Vizster: Searching 12

Vizster: Xray Mode 1 13

Vizster: Xray Mode 1 13

Vizster: Agglomerative Clustering 14

Vizster: Agglomerative Clustering 14

Vizster: Xray Mode 2 15

Vizster: Xray Mode 2 15

Node. Trix: A Hybrid Visualization of Social Networks (Henry, Fekete, Mc. Guffin) Overall layout

Node. Trix: A Hybrid Visualization of Social Networks (Henry, Fekete, Mc. Guffin) Overall layout is Lin. Log Energy. Based clustering By default node-link visualization Interaction used to select subgraph and transform into adjacency matrix Excellent for viewing clusters 16

Nodetrix 17

Nodetrix 17

Node. Trix: Bridge Identification 18

Node. Trix: Bridge Identification 18

Node. Trix: Cliques 19

Node. Trix: Cliques 19

Mat. Link: Enhanced Matrix Visualization for Analyzing Social Networks (Henry, Fekete) Matlink is a

Mat. Link: Enhanced Matrix Visualization for Analyzing Social Networks (Henry, Fekete) Matlink is a adjacency matrix enhanced with links for the current path Nodes are the “row/columns” Much easier to follow path Allows reordering of rows/columns: find cliques easier 20

Matlink: NL, MAT, Mat. Link 21

Matlink: NL, MAT, Mat. Link 21

Matlink 22

Matlink 22

Relation to Multi-agent systems agents are sociable need to cooperate and communicate in order

Relation to Multi-agent systems agents are sociable need to cooperate and communicate in order to meet a goal Graph visualizations are more scalable than a ladder visualization is messy for two to three agents but ladder is used because it expresses time well 23

Ladder vs. Node-Link Alice Bob what time is it? 5 o'clock Thanks 24

Ladder vs. Node-Link Alice Bob what time is it? 5 o'clock Thanks 24

 [1] J. Heer and D. Boyd. Vizster: Visualizing Online Social Networks. In Proceedings

[1] J. Heer and D. Boyd. Vizster: Visualizing Online Social Networks. In Proceedings of the 2005 IEEE Symposium on Information Visualization, pages 33– 40, 2005. [2] N. Henry and J. Fekete. Mat. Link: Enhanced Matrix Visualization for Analyzing Social Networks. LECTURE NOTES IN COMPUTER SCIENCE, 4663: 288, 2007. [3] N. Henry, J. D. Fekete, and M. J. Mc. Guffin. Node. Trix: A Hybrid Visualization of Social Networks. IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS, pages 1302– 1309, 2007. 25