A new Politico post produced by our lab examines the mentions of countries around the globe in congressional documents. We look at the frequency of country mentions and co-occurrence. These numbers were extracted from political statements provided by VoteSmart. Here we post the data we used for those visualizations.
The files below contain the frequency of country name occurrence and co-occurrence in Democrat, Republican, and total documents. The versions of the variables that have names ending in “n” are normalized and should be read as percent documents (Dem, Rep, or total) in which the terms occur. In the context of co-occurrence, the normalized weight of links between country A and country B is simply |A ∩ B| / |A|+|B| (the number of documents mentioning both A and B over the number of documents mentioning A plus the number of documents mentioning B).
Some of the more technically interesting plots I got to play with around this project were network maps using great circle arcs to connect geographic points. The R code for that will be posted here shortly.
Some additional charts that didn’t make it to the Politico post are included below.