LastFM Dataset

Description

This public dataset has one month of who-listens-to-which song information. We selected all 1000 users and the 1000 most listened songs resulting in 1,293,103 interactions. In this dataset, interactions do not have features.

Dataset Statistics

Users	Items	Interactions	Node Labels	Node Features	Edge Labels	Edge Features	Action Repetition (%)
980	1,000	1,293,103	None	None	None	None	8.6

References

Srijan Kumar, Xikun Zhang, and Jure Leskovec. 2019. Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD ’19). Association for Computing Machinery, New York, NY, USA, 1269–1278. DOI:https://doi.org/10.1145/3292500.3330895

 @inproceedings{kumar2019predicting,
	title={Predicting Dynamic Embedding Trajectory in Temporal Interaction Networks},
	author={Kumar, Srijan and Zhang, Xikun and Leskovec, Jure},
	booktitle={Proceedings of the 25th ACM SIGKDD international conference on Knowledge discovery and data mining},
	year={2019},
	organization={ACM}
 }

B. Hidasi and D. Tikk. Fast als-based tensor factorization for context-aware recommendation from implicit feedback. In ECML, 2012.

LastFM.tar.gz Download Link

Files

1. lastfm.tsv

• Tab-separated file Header: (row_id, user, item, timestamp)
• All user ids & item ids are normalized starting from 0.
• Sorted by timestamp.
• All NaN values are removed from data.
• Example

row_id	user_id	item_id	timestamp  
0	0	0	0  
1	1	1	54
2	1	2	306
3	2	3	479
......

1. lastfm_user_mapping.tsv & lastfm_item_mapping.tsv

• Mapping files from original user & item ids to normalized ones.
• Example

original_id	mapped_id
0	0
1	1
2	2
3	3
......

Contacts

Sejoon Oh, soh337@gatech.edu, Georgia Institute of Technology
Srijan Kumar, srijan@gatech.edu, Georgia Institute of Technology