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.
nlp text-mining twitter thesis open-data computational-social-science business-analytics political-polarization nova-sbe large-language-models congressional-record
-
Updated
Apr 8, 2026