Social Network Theory applied to Political Books

I continue to find social networks of information more interesting and informative than social networks of people. Perhaps this is because the richness of data available for mining on the Web of social information is so much greater. In five years, I can imagine that the existence of long data exhausts from Match.com, Friendster, political donation lists, Yahoo groups, etc., will create equally viable information sources for human social networks.

In the meantime, what we have to play with are social networks of information. Google, of course, is the exemplar of this field, with its PageRank algorithm based heavily on "what other sites think" of a given website. Amazon's 'people who bought X, also bought Y' is another fertile field; and we have a fascinating new analysis of political bestsellers from Valdis Krebs.

Here is the network map of the top 100 political books on Amazon, arranged according to their degree of relatedness. It demonstrates what Krebs rightly calls 'echo chambers' of thought on both the left and the right, with "debate replaced by hate" at the extreme margins, where partisans of both the left and right consistenly buy a tightly grouped set of books that most strongly reinforce each other.

This is a demonstration of the increasing segmentation of America, where sophisticated marketing, geographic mobility, and the twin philosophies of voter and consumer choice have enabled seemingly similar American citizens to exist in completely separate realities.

As an aside, I highly, highly recommend Charlie Wilson's War, which is one of the few books in this diagram read by both left and right. It's a superb story about rea-world people, and real-world politics, accomplishing literally unbelievable feats during the the closing years of the Cold War.