Exploring the characteristics of entire networks
In the next set of recipes, we will characterize our social network as a whole, rather than from the perspective of individual actors. This task is usually secondary to getting a feel of the most important nodes, but it is a chicken and an egg problem; determining the techniques to analyze and splitting the whole graph can be informed by key player analyses, and vice versa.
Getting ready
If you completed the previous recipes, you will be ready to proceed with this one.
How to do it...
The following steps will walk us through our first exploration of graph characteristics at the level of the whole graph:
- Let's compute both the density of the entire network and that of the ego graphs:
>>> nx.density(graph) 0.00810031232554 >>> ego = nx.ego_graph(graph, "LONGBOW/AMELIA GREER") >>> nx.density(ego) 0.721014492754
As you can see, our heroic social network is not very dense; it's not very cliquish as a whole. However, Longbow...