Quantifying Political Polarization Using Text Mining. Master's thesis at Nova School of Business and Economics. 5.6M Congressional speeches and tweets classified across five polarization domains by a four-model LLM ensemble.
-
Updated
Apr 8, 2026