Speaking of Refugees Discursive Power, Polarization and Framing in Europe’s Mediated Migration Debates - Project Data and Supplementary Materials

This repository serves as the central archival location for computational research materials, code, and supplementary data associated with the doctoral dissertation *Speaking of Refugees: Discursive Power, Polarization and Framing in Europe’s Mediated Migration Debates* (KU Leuven, Faculty of Social Sciences). The dissertation examines how migration and refugees are debated within Europe’s hybrid media system, where legacy journalism and social media publics interact and shape contemporary political communication. The empirical analyses address three major cases: (1) the 2015 European Refugee Crisis, (2) displacement following the 2022 Russian invasion of Ukraine, and (3) climate migration and refugee discourse on X (formerly Twitter). The data pertain to multiple languages, time periods, and media contexts related to these cases. Further details are provided in the dissertation and in the chapter- and file-level documentation included in this repository. Current contents Chapter-level folders corresponding to empirical studies in the dissertation; Python and R scripts and Jupyter notebooks documenting data collection, preprocessing, topic modeling (STM, BERTopic), social network analysis, and qualitative sampling workflows; Derived data products, including tweet ID datasets shared in accordance with X/Twitter Terms of Service; Static figures and interactive HTML visualizations accompanying the analyses; Chapter-specific and repository-level README files documenting analytical scope, data access constraints, and methodological decisions. Notes on updates This repository represents a curated snapshot of the computational materials underlying the dissertation at the time of submission. Minor updates may occur in the future to reflect accepted publications or finalized versions of manuscripts, but the overall structure and contents correspond to the dissertation as submitted. No full tweet text or user metadata will be shared, in line with ethical guidelines and platform data-sharing restrictions. Researchers may rehydrate tweet IDs through the X API depending on their institutional access policies. iCandid can be reached to request access to Flemish newspaper data.