This project analyzes the emotionality of political speeches within the EU Commission.

Individual speeches are scraped from the EU Commission Press Corner and matched with variables from the Chapel Hill Expert Survey. Word embeddings are calculated for each speech and for affective and cognitive wordlists from the LIWC dictionary. Emotionality is then derived from the cosine similarities of speeches with the dictionaries.

Emotionality in EU Commission topics are investigated across time, across political topics (using topic modeling), and by individual characteristics.