100,836 ratings and 3,683 tags applied to 9,742 movies by 610 users — the standard small recommender-systems dataset.
About
Five-star ratings and free-text tags from MovieLens, a movie recommendation service run by GroupLens at the University of Minnesota. Every user in the set rated at least 20 movies. This small version is meant for education and development; GroupLens publishes larger ones.
Files in the zip
| File | Rows |
|---|---|
ratings.csv |
100,836 — userId, movieId, rating, timestamp |
tags.csv |
3,683 |
movies.csv |
9,742 — title and pipe-separated genres |
links.csv |
IMDb and TMDb identifiers for each movie |
Licence
Not an open licence. You may not use the data for any commercial or revenue-bearing purpose without permission from GroupLens. You may redistribute it, including transformed versions, only under the same licence conditions. The full terms are in the README inside the zip. Please cite:
Harper, F. M., & Konstan, J. A. (2015). The MovieLens Datasets: History and Context. ACM Transactions on Interactive Intelligent Systems, 5(4). https://doi.org/10.1145/2827872