Datasets

DataRec includes several commonly used recommendation datasets to facilitate reproducibility and standardization. These datasets have been carefully curated, with traceable sources and versioning information maintained whenever possible. For each dataset, DataRec provides metadata such as the number of users, items, and interactions and data characteristics known to impact recommendation performance (e.g., sparsity and user/item distribution shifts). The dataset collection in DataRec is continuously updated to include more recent and widely used datasets from the recommendation systems literature. The most recent and widely used version is included when the original data source is unavailable to ensure backward compatibility.

The following datasets are currently included in DataRec:

Dataset	Version	Dataset Page
Alibaba iFashion	v1	page
Amazon Baby	2023	page
Amazon Beauty	2023	page
Amazon Books	2023	page
Amazon Clothing	2023	page
Amazon Music	2023	page
Amazon Office	2023	page
Amazon Sports and Outdoors	2023	page
Amazon Toys and Games	2023	page
Amazon Video Games	2023	page
Ambar	2024	page
CiaoDVD	v1	page
CiteULike	a	page
	t	page
Epinions	v1	page
Gowalla	checkins	page
	friendships	page
LastFM	2011	page
Mind	large	page
	small	page
Movielens	100k	page
	1m	page
	20m	page
Tmall	v1	page
Yelp	v1	page