
Analogous to the co-authorship graphs that have been constructed to show collaboration patterns within academic communities, one could build general common-interest graphs that show website networks for a given topic. A first example could be the sphere of netlabels. Tools would be: google search (via API), maybe a custom web crawler, and a graph visualization toolkit.
Tracking graph changes over time and finding a meaningful way of visualizing temporal differences would be an interesting research topic.
Some relevant links:
http://www.research.att.com/sw/tools/graphviz/
AT&T Graph Visualization Library
http://www.research.att.com/areas/visualization/
AT&T Information Visualization Group
http://www.aisee.com
Graph Visualization from Saarbruecken, Germany
http://tomsawyer.com/home/index.php
Another graph visualization software company. Used by Alias for Maya.
Tangentially related:
http://www2.parc.com/istl/projects/uir/
Xerox Parc UI Group