9.7. Visualizing Social Networks

As we saw in Section Social Networks, the problem of visualizing social networks starts out with the problem of laying out the network for display. This involves a number of heuristics such as discovering symmetries, bringing together nodes with many cinnections, minimizing tehnumber of crossing lines, and so on.

Zachary’s karate club network: View 1

Zachary’s karate club network: View 2

9.7.1. Combining hierarchical information with other kinds of links

Details of the construction of the following graph of the Enron email network are given in [SHETTY2005] .

Enron email network. A link means that the participants exchanged email (at least one message in each direction) and that at least 5 messages were sent.

A key fact about the Enron data is that it represents email exchanges between individuals in a organization. Therefore, there is a pre-existing hierarchical social structure roughly captured by the organization tree. The layout of the nodes in the graph shown above incorporates that information using multidimensional scaling (MDS), representing closeness in the organizational hierachy with closeness in the graph.