YouTube Radicalization (2008 - 2018)

This dataset was created by Dutch journalists at De Correspondent and De Volkskrant in order to investigate YouTube Radicalization. Their work is detailed in various Dutch articles and presented as a case-study in the Global Network of Investigative Journalism in English. In their work they describe that the dataset contains approximatetly: 

600,000 videos

450,000 transcripts of those videos (by using YouTube’s automatic closed-captioning service, which is not available for all videos)

120 million comments on those videos

20 million recommendations automatically generated from viewing those videos

The dataset has also been used for academic research at the department of Mediastudies at the University of Amsterdam. Notably for the publication of 'A free market in extreme speech: Scientific racism and bloodsports on YouTube' in the journal Digital Scholarship in the Humanties. The dataset is managed by Ivan Kisjes as part of the Public Data Lab.

As of October 2022, the dataset is also used by Daniel Jurg, who now manages this version of the dataset for his PhD project: 'Broadcasting the Culture War - Audience Engagement in YouTube’s Alternative Community,' as funded by the FWO (FWOTM1133). In the final SQL version of the dataset that is captured in this repository, there is the following data:

channels: 1738

videos: 626110

comments: 109,315,583

Given the sensitivity of this data it cannot yet be widely shared and we are looking working on ways to acrhive a public version.