Exploring Large Social Networks with Matrix-Based Representations... Ph.D. Thesis

social networks are groups of persons and their relations. A common example is a genealogical tree or a telecommunication (phone, email) network.

Social Networks 1 [www] How to create a network? How to explore it? Three groups of social science researchers brainstormed on these questions.

Social Networks 2 [www] How to explore a network? How to communicate findings? Three groups of social science researchers brainstormed on these questions.

matrices represent social networks as tables containing actors (persons) both in rows and columns and marking their relations (connections) in intersecting cells. I investigated the potential of these representations in my Ph. D.

MatrixExplorer [video] [pdf] Matrix representations are good for dense networks, node-link diagrams for sparse one. Combine best of both worlds with MatrixExplorer.

MatLink [video] [pdf] Following paths on matrices is a tedious task. MatLink augment standard matrices with static and interactive links to help perform this family of tasks.

Mélange [video] [pdf] Want to fold your screen as a piece of paper? Melange is an interaction technique that folds the space for multiple foci while preserving the intermediate context.

NodeTrix [video] [pdf] Small-world networks are globally sparse but locally dense. Node-link diagrams are good for sparse networks, matrices are good for dense ones. Let's merge them with NodeTrix!

Node Duplication [pdf] Where to place people that belong to two or more communities? Let's duplicate them and see how it affects user understanding!

exploring publication data

UIST poster [pdf] [data] 20 years of collaboration at UIST. Coauthorship network for long papers and notes: a 1m x 2m poster.

HCI Publication Data [pdf] GraphML Data: [Infovis] [UIST] [AVI] [CHI (5Mo)] [Infovis+UIST+AVI+CHI (7Mo)] How to become a successful researcher? Who are the central actors in HCI? Who published the most over 25 years of CHI conferences? Visual answers to these questions.

very large matrices

ZAME [pdf] Exploring links between 6 millions of pages of the french wikipedia sounds like something impossible? Check out ZAME, the Zoomable Adjacency Matrix Explorer designed by the AVIZ team!