MGPHot: A Dataset of Musicological Annotations for Popular Music (1958-2022)

The MGPHot Dataset includes more than 21k~songs that have appeared at least once in the Billboard Hot 100 charts from 1958 until 2022, annotated by professional musicologists with 58 musical attributes which are grouped in seven different categories: Rhythm, Compositional Focus, Harmony, Instrumentation, Sonority, Vocals and Lyrics. This dataset has been initially built and used for the study described in the paper referred to below, on the evolution of popular music over the past 65 years. It is however not restricted to it, as we believe is opens up many academic endeavors in Musicology and Music Information Retrieval (e.g., music description modelling, auto-tagging, chart prediction, music recommendation, etc.). Overview of files: mgphot_genes.tsvContains the list of genes (i.e., musical attributes), each accompanied by a description. The index of each gene is important, as it is used to identify corresponding values in mgphot_gene_values.tsv. mgphot_gene_values.tsvProvides the gene values for all tracks in the dataset, along with metadata such as the year of first appearance on the Billboard charts, the song title, and the artist (as listed in the Billboard dataset). Each track is also assigned a unique ID (mgphot_track_id), which is used to match entries in hot100_charts.tsv. hot100_charts.tsvContains the Billboard chart data used in the analysis, filtered to include only the songs present in the MGPHot dataset. The mgphot_track_id field allows matching these entries with the tracks listed in mgphot_gene_values.tsv. How to use: If you want to use this dataset, please first consult the license, and make sure the terms are acceptable to you. Make sure to provide the following citation in your work: Oramas, S., Gouyon, F., Hogan, S., Landau, C., & Ehmann, A. (2025). MGPHot: A Dataset of Musicological Annotations for Popular Music (1958–2022). Transactions of the International Society for Music Information Retrieval, 8(1), 1–13. DOI: https://doi.org/10.5334/tismir.236