PolitiFact scraping dataset

This dataset comprises a comprehensive and continuously updatable collection of fact-checked political statements sourced from PolitiFact.

We provide the code for a custom scraping tool that downloads the full historical content of PolitiFact, including article reviews, associated speakers, reviewers, and topical issues. This allows authors do automatically download updated versions of this dataset. The code provided allows to scrape the information and, additionally, store it in a MongoDB database. This last step is optional; the data included in this Zenodo entry corresponds to the MongoDB data but can be used without storing it in a MongoDB database.

Each entry includes the original statement, metadata (e.g., dates, titles), detailed review analyses, cited sources, and a veracity label based on the six-level Truth-O-Meter scale (from “True” to “Pants on Fire”). Additional enriched information is provided for speakers (e.g., descriptions, activity statistics) and reviewers (e.g., roles, profiles), as well as issue-level categorization.