The Pushshift Telegram Dataset

  • Jason Baumgartner Pushshift.io
  • Savvas Zannettou Max-Planck-Institut für Informatik
  • Megan Squire Elon University
  • Jeremy Blackburn Binghamton University

Abstract

Messaging platforms, especially those with a mobile focus, have become increasingly ubiquitous in society. These mobile messaging platforms can have deceivingly large user bases, and in addition to being a way for people to stay in touch, are often used to organize social movements, as well as a place for extremists to congregate.

In this paper, we present a dataset from one such mobile messaging platform: Telegram. Our dataset is made up of over 27.8K channels and 317M messages from 2.2M unique users. To the best of our knowledge, our dataset comprises the largest and most complete of its kind. In addition to the raw data, we also provide the source code used to collect it, allowing researchers to run their own data collection instance. We believe the Pushshift Telegram dataset can help researchers from a variety of disciplines interested in studying online social movements, protests, political extremism, and disinformation.

Published
2020-05-26
How to Cite
Baumgartner, J., Zannettou, S., Squire, M., & Blackburn, J. (2020). The Pushshift Telegram Dataset. Proceedings of the International AAAI Conference on Web and Social Media, 14(1), 840-847. Retrieved from https://www.aaai.org/ojs/index.php/ICWSM/article/view/7348