A Frustratingly Easy Way of Extracting Political Networks from Text

The dataset accompanying the paper "A Frustratingly Easy Way of Extracting Political Networks from Text" contains data used to validate the extraction of political networks from Chilean political news articles using the GPT-4 language model. The dataset includes a corpus of 1009 news articles mentioning members of the Chilean Chamber of Deputies, metadata on the relationships identified between deputies, and sentiment analysis results. The experimental procedures involve linear regression and node embedding analyses to establish correlations between identified sentiments and legislative voting patterns among deputies.