Nice! Do you know if Gephi can do something similar to the summarization that we did using cluster diagrams? The whole "ball of hair" problem doesn't have any other real solution, I don't think (well, unless you use edge clustering, but that doesn't help in-group connections).