German Parliamentary Speeches

<div> <div>These datasets are part of the thesis entitled "A natural language processing analysis of parliamentary speeches in the German Bundestag from 1949 to 2023 with regards to gender equality and women's politics". The focus of the thesis lies on contrasting the thematic preferences of men and women, as well as their connection to women's political issues based on their speeches and highlighting their historical development. Furthermore, the speakers&rsquo; reception in parliament is analyzed and the gender-specific distinguishability of speech style is verified. The data is available in a pandas DataFrame format.</div> <div>&nbsp;</div> <div> <div> <div>The plenary protocol data has been extracted from the following two sources:</div> <div> <ul> <li>Blaette, Andreas (2017): GermaParl. Corpus of Plenary Protocols of the German Bundestag. TEI files, availables at:&nbsp;<a href="https://github.com/PolMine/GermaParlTEI">https://github.com/PolMine/GermaParlTEI</a></li> </ul> </div> <div> <ul> <li>Deutscher Bundestag. Open Data - Plenarprotokolle der 20. Wahlperiode und Stammdaten aller Abgeordneten seit 1949.&nbsp;<a href="https://www.bundestag.de/services/opendata">https://www.bundestag.de/services/opendata</a></li> </ul> </div> <div>&nbsp;</div> <div>The following datasets are contained:</div> <div> <ul> <li><strong>data_merged_revised.pkl:</strong> DataFrame containing the extracted and cleaned data from the sources mentioned above&nbsp;</li> <li><strong>data_topics_revised.pkl</strong>: DataFrame containing the data from the sources mentioned above and the topic distributions after topic modelling</li> <li><strong>data_undersampled_processed.pkl</strong>: Undersampled and preprocessed dataset which can be used for training classification models</li> </ul> </div> </div> </div> </div>