It is a commonplace among political historians and the punditocracy to discuss the polarization of the body politic in societies at various moments. Well, here’s an example of how digital media can be used to visualize the polarization of the American body politic in 2004. Valdis Krebs of orgnet.com (Warning: slow loading site with large image files!) used InFlow software to suck down into a database information on books purchased from major web book retailers. Krebs networked this information by linking books if they were purchased by the same person — “Customers who bought this book also bought…”
The resulting social network of books demonstrates just how much Americans buying the top 100 political books of 2004 were living in completely different worlds. Only 15 of the top 100 books in Krebs’ database had some sort of cross-over appeal from left to right. Mostly, those buying leftish political books bought more leftish political books and those buying rightish books bought other rightish books. It’s impossible to say, of course, whether these purchase patterns were a symptom or a cause of the polarization of American political culture, but the social network map of these purchases certainly shows that polarization.
What other data might historians begin using to come up with similar visualizations of the past? Right now we are largely limited to socioeconomic data. But, for instance, what might one be able to learn by ingesting stories from the New York Times over many years and analyzing the ways that those stories describe, for instance, lynchings in the American South, or anti-Semitism in Central Europe? When it comes to thinking carefully about how humanists should begin to approach computational humanities of this sort, be sure to read Dan Cohen’s most recent musings on this subject.