PRIPiT: An Annotated Dataset for Propaganda Detection through Ideology, Power Relations and Persuasion in Politicians' War-related Tweets

The dataset consists of 1000 English-language tweets posted by Russian and Ukrainian politicians during the ongoing war. These tweets were annotated with 45 distinct labels divided into categories like genre, ideologies, power relations and persuasion. The dataset is developed for dual purposes; the first one is to analyse the propaganda by integrating critical discourse analysis (CDA) research. On the other hand, the dataset supports multi-label classification research in computational propaganda detection. The dataset is designed for researchers, NLP practitioners, and educators seeking to explore fine-grained propaganda detection in social media. Its goal is to contribute to more transparent, fair, and effective models in digital media analysis with short texts.