A cluster analysis of harmony in the McGill Billboard dataset. Shaffer, K., Vasiete, E., Jacquez, B., Davis, A., Escalante, D., Hicks, C., McCann, J., Noufi, C., & Salminen, P. Empirical Musicology Review, 14(3-4):146, jul, 2020.
A cluster analysis of harmony in the McGill Billboard dataset [link]Paper  doi  abstract   bibtex   
We set out to perform a cluster analysis of harmonic structures (specifically, chord-to-chord transitions) in the McGill Billboard dataset, to determine whether there is evidence of multiple harmonic grammars and practices in the corpus, and if so, what the optimal division of songs, according to those harmonic grammars, is. We define optimal as providing meaningful, specific information about the harmonic practices of songs in the cluster, but being general enough to be used as a guide to songwriting and predictive listening. We test two hypotheses in our cluster analysis — first that 5–9 clusters would be optimal, based on the work of Walter Everett (2004), and second that 15 clusters would be optimal, based on a set of user-generated genre tags reported by Hendrik Schreiber (2015). We subjected the harmonic structures for each song in the corpus to a K-means cluster analysis. We conclude that the optimal clustering solution is likely to be within the 5–8 cluster range. We also propose that a map of cluster types emerging as the number of clusters increases from one to eight constitutes a greater aid to our understanding of how various harmonic practices, styles, and sub-styles comprise the McGill Billboard dataset.
@Article{          shaffer.ea2020-cluster,
    author       = {Shaffer, Kris and Vasiete, Esther and Jacquez, Brandon
                   and Davis, Aaron and Escalante, Diego and Hicks, Calvin
                   and McCann, Joshua and Noufi, Camille and Salminen, Paul},
    year         = {2020},
    title        = {A cluster analysis of harmony in the McGill Billboard
                   dataset},
    abstract     = {We set out to perform a cluster analysis of harmonic
                   structures (specifically, chord-to-chord transitions) in
                   the McGill Billboard dataset, to determine whether there
                   is evidence of multiple harmonic grammars and practices in
                   the corpus, and if so, what the optimal division of songs,
                   according to those harmonic grammars, is. We define
                   optimal as providing meaningful, specific information
                   about the harmonic practices of songs in the cluster, but
                   being general enough to be used as a guide to songwriting
                   and predictive listening. We test two hypotheses in our
                   cluster analysis — first that 5–9 clusters would be
                   optimal, based on the work of Walter Everett (2004), and
                   second that 15 clusters would be optimal, based on a set
                   of user-generated genre tags reported by Hendrik Schreiber
                   (2015). We subjected the harmonic structures for each song
                   in the corpus to a K-means cluster analysis. We conclude
                   that the optimal clustering solution is likely to be
                   within the 5--8 cluster range. We also propose that a map
                   of cluster types emerging as the number of clusters
                   increases from one to eight constitutes a greater aid to
                   our understanding of how various harmonic practices,
                   styles, and sub-styles comprise the McGill Billboard
                   dataset.},
    doi          = {10.18061/emr.v14i3-4.5576},
    issn         = {1559-5749},
    journal      = {Empirical Musicology Review},
    keywords     = {McGill Billboard dataset,cluster analysis,harmonic
                   syntax,machine learning,music analysis with
                   computers,pop/rock,rock,transitional
                   probability,visualization},
    mendeley-tags= {music analysis with computers},
    month        = {jul},
    number       = {3-4},
    pages        = {146},
    url          = {https://emusicology.org/article/view/5576},
    volume       = {14}
}

Downloads: 0