Example layouts of the edge-repulsion LinLog energy model [3,4]:
In all pictures, nodes are represented as circles, with an area proportional to the degree (the total edge weight) of the node. In most pictures, edges are elided to avoid clutter.
For all examples, the drawings of the edge-repulsion LinLog energy model [3,4] are compared with drawings of the node-repulsion LinLog energy model [2,4] and the well-known Fruchterman-Reingold force model [1]. In general, both node-repulsion energy models tend to group nodes with high degree, while the edge-repulsion LinLog model shows more interesting groups. Both LinLog models tend to separate clusters better than the Fruchterman-Reingold model.
For medium and large graphs, VRML files are provided because they enable navigation and selective showing of node labels. Viewing these files requires a VRML viewer, for example Cortona VRML Client, a plugin for standard web browsers.
The tool LinLogLayout with reusable Java classes (LGPL license) for computing LinLog drawings and Modularity clusterings is available here.
[1] Thomas M. J. Fruchterman and Edward M. Reingold: Graph Drawing by Force-Directed Placement. Software - Practice and Experience 21(11): 1129-1164, 1991
[2] Andreas Noack. An Energy Model for Visual Graph Clustering. In Proceedings of the 11th International Symposium on Graph Drawing (GD 2003), LNCS 2912, pages 425-436. © Springer-Verlag, 2004. (Abstract, PDF, Publisher)
[3] Andreas Noack. Energy-Based Clustering of Graphs with Nonuniform Degrees. In Proceedings of the 13th International Symposium on Graph Drawing (GD 2005), LNCS 3843, pages 309-320. © Springer-Verlag, 2006. (Abstract, PDF, Publisher)
[4] Andreas Noack. Energy Models for Graph Clustering. Journal of Graph Algorithms and Applications 11(2):453-480, 2007. (Abstract, PDF, Publisher)
Author: Andreas Noack