Preferred algorithms for exploration?

[Anthony Moser]

Hey!

I've been making a set of tools for generating network graphs of public records data that are built on networkx.

Right now there are three main ones:

• graphemails.com visualizes email logs
• deseguys.com visualizes business records, campaign contributions, and contracts
• bit.ly/qng accepts arbitrary spreadsheets and lets you choose columns to link together then generates a graph out of them

I've used the simple paths option to highlight how nodes are connected, and community detection for updating the node colors, and I've tried varying the size of nodes based on degree / centrality. 

But networkx has a lot of algorithms for detecting cycles, shortest paths, etc, and I'd like to know if folks have suggestions for what might be useful approaches for exploring what are mostly social networks. 

I know there's a lot of literature out there but I wanted to see if folks have practical pointers or go-to methods, especially for simplifying, identifying key points, or drawing inferences about the nature of the graph from its structure.

Thank you!

Anthony Moser