Fact-Checked Facebook News Corpus from Bangladesh

Published: 15 December 2023| Version 3 | DOI: 10.17632/ckkpwwydw8.3
, Musa Akram Saleh, Abdul Ali Hayder Khan, Md Abu Rayhan, Khandaker Foysal,


This dataset is a meticulously curated collection of Bengali fact-checked fake news reports, assembled by the FactWatch team, a renowned IFCN signatory and Meta's third-party fact-checking partner in Bangladesh. Spanning from October 4, 2021, to May 25, 2023, it comprises 1014 comprehensive reports that encapsulate the rise of misinformation during the Covid-19 pandemic, particularly on Facebook. Each report in this dataset has been manually annotated by expert annotators from the FactWatch team. The annotations include detailed summaries, report categories, accurate veracity labels, examples of the original fake news content, and in-depth evaluations of the fact-checking process. This labor-intensive annotation process ensures high-quality, reliable data. The dataset stands as a critical resource for research in the domains of misinformation analysis, natural language processing, and automated fact-checking systems. It is particularly valuable for those focusing on the Bengali language and information ecosystem, providing a foundational dataset for further study and innovation in combating digital misinformation.



East-West University, University of Liberal Arts Bangladesh


Social Media, Natural Language Processing, Facebook, Misinformation Effect