Discovering a Shared Past: Topic Modelling of Austrian Historical Newspapers

This study examines topics in Austrian historical newspapers published in the second half of the 19th century, specifically “Das Vaterland” and “Neue Freie Presse”, which are part of the ANNO collection (Austrian National Library, 2021). The primary aim is to identify and analyse topics related to national minorities, labour, and migration in the multi-ethnic Habsburg Empire. The study outlines the process of data collection, pre-processing, and Dynamic Topic Modelling (DTM) using BERTopic. The results demonstrate that BERTopic, enriched with seed words, successfully identifies topics related to migration, national minorities, labour, and nationalistic movements. DTM revealed that historical events could be identified through the frequency of these topics at specific points in time, highlighting its effectiveness in the topic-specific corpus-building process. These results will serve as a foundation for constructing a migration corpus of Austrian newspapers, enabling further in-depth analyses in the future.