- Sergio Camelo
Graphs are a great mechanism to visualize relationships between objects. Think of your social graph on Facebook, academic collaboration networks or economics networks. Graphs are well understood mathematical objects, but there does not seem to exist a "correct" way to embed a graph in the plane. My project, then, is to explore different ways to visualize graphs. Hopefully, this will reveal some hidden structure, or make a concept like "centrality" more easy to understand.
Most of the techniques used for graph visualization are "force" algorithms, in which the edges of the graph act as springs and the final visualization ends up being the equilibrium state of the system. This gives very good results, but does not take into account centrality or power measures of the nodes in the graph. Many times, the user wants to determine what are the most important nodes of the graph, either because they are a node through which other nodes connect, or because they have a privileged position to exercise some kind of control over less powerful nodes. "Force" algorithms do not reveal this structure, but if we integrate centrality measures into the springs' hook constant, or as a gravity force that attracts or repel nodes, we may get much better visualizations. For my project, I want to combine this idea with the d3-force library.
Project Progress Presentation
- Link to source code and executable
- Link to final paper in pdf form
- Link to final slides or poster