Analyzing the WikiLeaks Data
| Peter Klein |
Once more on WikiLeaks: A team of University of Colorado researchers has already produced a geospatial analysis of the incident reports contained in the dataset. “By mapping the violence and examining its temporal dimensions, the authors explain its diffusion from traditional foci along the border between the two countries. While violence is still overwhelmingly concentrated in the Pashtun regions in both countries, recent policy shifts by the American and Pakistani governments in the conduct of the war are reflected in a sizeable increase in overall violence and its geographic spread to key cities. . . .” This is exactly the kind of analysis the military intelligence agencies are not doing, or at least not sharing.
Economists, geographers, entrepreneurship and innovation researchers, and other social scientists have a lot of expertise in network and cluster analysis. Why not turn them loose on these kinds of raw data? It’s also cheap: as Karen Kwiatkowski notes, “[t]he study was honestly, scientifically, and nimbly completed and published at no direct cost to the intelligence community. It was made possible by the decentralization, fluidity, and constant sharing and shifting of roles and responsibilities that comprise the Internet.”