Schaal, MarkusDavis, EnnoMueller, Roland M.2021-12-242021-12-242022-01-04978-0-9981331-5-7http://hdl.handle.net/10125/79700In recent years, automated political text processing became an indispensable requirement for providing automatic access to political debate. During the Covid-19 worldwide pandemic, this need became visible not only in social sciences but also in public opinion. We provide a path to operationalize this need in a multi-lingual topic-oriented manner. Using a publicly available data set consisting of parliamentary speeches, we create a novel process pipeline to identify a good reference model and to link national topics to the cross-national topics. We use design science research to create this process pipeline as an artifact.10 pagesengAttribution-NonCommercial-NoDerivatives 4.0 InternationalData Analytics, Data Mining and Machine Learning for Social Mediacross-countrylatent dirichlet allocationmulti-lingualparliamentary speechprobabilistic topic modellingtext10.24251/HICSS.2022.366