Social Network Analysis Isn't Just For People

The tools of social network analysis – centrality measures, clustering, graph-traversal algorithms, community detection and so forth – are largely based on mathematical network theory. There is very little in these techniques that actually requires that the data represents social activity. This presentation will show how these techniques can be applied to data from areas such as geo, the Wikipedia link graph and linguistics.

We’ll show how to take tabular or textual data and derive graph representations from it that can be used to apply these techniques. We’ll discuss practical applications of these techniques in delivering new features for web applications. We’ll also show how the powerful visualisation tool Gephi can be used to explore the data once it’s in graph form.

This talk will be partly based on content from an Ignite talk given at Strata NYC 2011: http://slideshare.net/mattb/place-graphs-are-the-new-social-graphs