This data set contains 1,019,318 unique users' music play counts in the Echo Nest, which is available at "http://millionsongdataset.com/tasteprofile/". As a basic step, it is interesting to predict the play counts using the song information collected in the Million Song Dataset (Bertin-Mahieux et al. (2011)). After cleaning up and feature engineering the data in total contains 205,032 observations where we consider the covariates duration, loudness, tempo, artist hotness, song hotness, and album hotness to model the response, the number of song counts.
Format
A data frame with 4 columns and 309,685 rows.
CountsNumber of playback counts for songs
DurationDuration of the song
LoudnessLoudness of the song
TempoTempo of the song
Artist_HotnessA value between 0 and 1
Song_HotnessA value between 0 and 1
Album_HotnessA value between 0 and 1
References
McFee B, Bertin-Mahieux T, Ellis DP, Lanckriet GR (2012). “The million song dataset challenge.” In Proceedings of the 21st International Conference on World Wide Web, 909--916. Ai M, Yu J, Zhang H, Wang H (2021). “Optimal subsampling algorithms for big data regressions.” Statistica Sinica, 31(2), 749--772.
Examples
nrow(One_Million_Songs)
#> [1] 205032