FUBU-Epstein: A Large-Scale Twitter Dataset on the Jeffrey Epstein Case and Its Global Public Discourse

Paper here https://arxiv.org/abs/2608.10210 FUBU-Epstein is a textless, anonymized derivative of a large Twitter/X research corpus on public discourse around the Jeffrey Epstein case. The release is designed for reproducible aggregate and network analysis while excluding tweet text and original platform identifiers. The package contains 54,375,364 status rows, 13,983,232 unique-text annotation rows, and 46,150,021 retained internal retweet, reply, and quote relationships. Status rows can be joined to annotations by annotation_id, and retained relationships can be joined to statuses by randomized record_id values. The public release excludes original tweet/status IDs, original user IDs, tweet text, handles, mentions, hashtags, URLs, locations, profile fields, exact timestamps, reverse lookup maps, randomization seeds, and relationships whose target status is outside the release. User, status, and annotation identifiers are release-only randomized contiguous integer IDs. Time is represented through a total chronological_index, created_month for all rows, and created_day for the main 2019-08-03 to 2023-04-03 collection period. Exact timestamps are not included. Annotations include sentiment, conspiracy, misinformation, toxicity, moral-emotion, ideology, partisanship, mobilization, persuasive-framing, bot/spam-likelihood, and language fields, plus selected aggregate summaries. The archive includes a README, data dictionary, manifest, validation report, checksums, parquet data files, and aggregate CSV files.To request access to raw data for collaborative research under ethical and legal safeguards, contact [e-mail] When using the dataset for your work. Please cite @misc{kreil2026fubuepsteinlargescaletwitterdataset, title={FUBU-EPSTEIN: A Large-Scale Twitter Dataset on the Jeffrey Epstein Case and Its Global Public Discourse (2019-2023)}, author={Michael Kreil and Tristan Manfred Stöber and Daniel Thilo Schroeder}, year={2026}, eprint={2608.10210}, archivePrefix={arXiv}, primaryClass={cs.SI}, url={https://arxiv.org/abs/2608.10210}, }