Dataset for the paper - Bots, Elections, and Controversies: Twitter Insights from Brazil's Polarised Elections

Dataset for the paper "Diogo Pacheco. 2024. Bots, Elections, and Controversies: Twitter Insightsfrom Brazil’s Polarised Elections. In Proceedings of the ACM Web Conference2024 (WWW ’24), May 13–17, 2024, Singapore, Singapore. ACM, New York,NY, USA. https://doi.org/10.1145/3589334.3645651" (arXiv version - https://doi.org/10.48550/arXiv.2310.09051).

The dataset spans from August 30, 2018, to March 14, 2023. The period encompasses 1,657 days, and the collection process remained active for 94% of this time. This comprehensive effort resulted in the acquisition of a vast dataset comprising 437 million tweets originating from 13 million distinct accounts.

The dataset is composed of CSV files, grouped by year. Each record represents tweets and has: masked 'id_str', masked  'user.id_str', 'timestamp_ms', masked 'retweeted_status.id_str', masked 'quoted_status.id_str', masked  'in_reply_to_status_id_str', and 'botscore' (given by BotometerLite). An example file is given for easy reference.

For more details about the data collection, please refer to the paper.