Adventures in Music Analytics
Dominique Haughton, Bentley University, Université Paris 1 and Université Toulouse 1, with the Music Analytics Group at Bentley University
This talk will give an overview of projects at the frontier between data analysis and music (“music analytics”) and will present preliminary results of an analysis of the success of Kickstarter rock music projects, using features extracted from audio files with the Echonest/Spotify API (Application Program Interface). We also discuss current state of the art in defining measures of distance between music excerpts and potential applications to a classification of jingles (with a view to understanding what makes a jingle successful, and indeed how one might measure success for a jingle) and to unraveling legal issues such as claims of plagiarism (as for example in Led Zeppelin’s song Stairway to Heaven). The talk will conclude with a brief mention of other potential areas of interest, such as the automatic identification of allegories in music and approaches to a classification of Chinese classical music. The discussion relies on work conducted by the recently formed music analytics group at Bentley University.