FACTurk: A New Dataset for Studying the Turkish Fact-checking Ecosystem

This repository includes the dataset and source codes presented in the paper titled "FACTurk: Insights into the Turkish Fact-Checking Ecosystem via a New Dataset" accepted for the Computational Social Sciences Special Session at the 34th Conference on Signal Processing and Communications Applications (SIU 2026).

FACTurk contains 15,766 fact-checking reports published in 2009-2026 and collected from three IFCN-accredited (Teyit.org, Doğruluk Payı, and Doğrula) and three other (Malumatfuruş, Yalansavar, and Günün Yalanları) Turkish fact-checking organisations. More details are available on the GitHub page

Dataset Structure

FACTurk dataset is provided in CSV format, which contains the following columns:

Field Name	Description
Row ID	Primary Key
Report ID	ID given to each unique report
Claim ID (High conf.)	ID given to each unique claim (with 0.95 similarity threshold)
Claim ID (Low conf.)	ID given to each unique claim (with 0.88 similarity threshold)
Claim	Textual claim being fact-checked
Content	(not published to prevent copyright infringement)
Date published	Date of publication of the report in ISO 8601 format
Author	Author(s) of the fact-checking report
Organisation	Name of the fact-checking organisation publishing the report
Original verdict	Conclusion of the fact-check
Title	Heading of the report
URL	Online link to the report
Normalised rating	One of predefined ratings derived from original verdict

Contact

If you have any questions about the dataset and/or source codes, please reach out to the contributor via email: Enes Altuncu

Citation

Please cite this work as follows if you use FACTurk or the provided source codes in your research:

@inproceedings{FACTurk2026,
  title={{FACTurk}: Insights into the Turkish Fact-Checking Ecosystem via a New Dataset},
  authors={Altuncu, Enes},
  year={2026},
  numpages={4},
  doi={},
  booktitle={34th Conference on Signal Processing and Communications Applications (SIU '26)},
  publisher={IEEE},
}