aft4.tsv.gz

1-year dump of Wikipedia article ratings. The dataset includes 47,207,448 records corresponding to 11,801,862 unique ratings posted between July 22, 2011 and July 22, 2012. A small sample is available as a separate file.

The following fields are included:

1. timestamp: 14-digit timestamp (UTC)
2. page_id: article ID
3. page_title: full article title
4. page_namespace: article namespace (0 is for Wikipedia's main article namespace)
5. rev_id: revision ID (article version that was rated)
6. user_id: boolean field set to 0 for anonymous raters and to 1 for registered users
7. rating_key: numeric key of the rating dimension (1: trustworthy, 2: objective, 3: complete, 4: well-written)
8. rating_value: rating value on a 1-5 scale. A value set to 0 indicates that the rater didn't rate the article on the corresponding dimension.