A project analyzing and predicting falloff in news viewership between original stories and updates/retractions.

This dataset, scraped from a variety of news media, including print and video, measures, where possible, how many viewers have seen an article or news story. Where this data is not available, it measures comparative audience interaction (e.g., through number of comments or social media shares).

The aim of this data is observe the difference in vewership and interaction between an original news story and a correction or retraction of that same news story. It also aims to study the snowball effect of that incorrect or un-updated piece of content and how it is shared with other news outlets and how often the corrected version travels to the same sources.