Investigating Cultural Stratification through Music Listening Data: An Initial Exploration with Machine Learning

Relying on a mixed-method survey design, we jointly analyze, at the individual level, observational data on music listening histories from streaming platforms alongside sociodemographic data from questionnaire surveys. This approach allows us to capture an unprecedented level of granularity and precision in the music listening habits of over 15,000 respondents over a period of more than five years, while connecting these observational data to the individuals' sociodemographic characteristics. We explore the potential of machine learning methods to revisit the relationships between social positioning variables and cultural practices. Our goal is to critically assess models of distinction, omnivorousness, and generational differentiation, in order to further examine contemporary forms of cultural stratification in the digital age.