Cultural differences in music features across Taiwanese, Japanese and American markets

Abstract

Background Preferences for music can be represented through music features. The widespread prevalence of music streaming has allowed for music feature information to be consolidated by service providers like Spotify. In this paper, we demonstrate that machine learning classification on cultural market membership (Taiwanese, Japanese, American) by music features reveals variations in popular music across these markets. Methods We present an exploratory analysis of 1.08 million songs centred on Taiwanese, Japanese and American markets. We use both multiclass classification models (Gradient Boosted Decision Trees (GBDT) and Multilayer Perceptron (MLP)), and binary classification models, and interpret their results using variable importance measures and Partial Dependence Plots. To ensure the reliability of our interpretations, we conducted a follow-up study comparing Top-50 playlists from Taiwan, Japan, and the US on identified variables of importance. Results The multiclass models achieved moderate classification accuracy (GBDT = 0.69, MLP = 0.66). Accuracy scores for binary classification models ranged between 0.71 to 0.81. Model interpretation revealed music features of greatest importance: Overall, popular music in Taiwan was characterised by high acousticness, American music was characterised by high speechiness, and Japanese music was characterised by high energy features. A follow-up study using Top-50 charts found similarly significant differences between cultures for these three features. Conclusion We demonstrate that machine learning can reveal both the magnitude of differences in music preference across Taiwanese, Japanese, and American markets, and where these preferences are different. While this paper is limited to Spotify data, it underscores the potential contribution of machine learning in exploratory approaches to research on cultural differences.

Energetic music is used for anger downregulation: A cross‐cultural differentiation of intensity from rhythmic arousal

“…This should then be reflected in the arousal markers (i.e., intensity: energy, rhythm: danceability) from that culture's charts. For danceability, a rhythmic marker, this effect was notably consistent, in that countries with a stronger prevalence of negative emotions, and loose (as opposed to tight) social norms had higher danceability in their top‐50 charts (Liew et al, 2021a, 2021b). Yet, in this analysis, energy, while consistently predicted by negative emotion prevalence, was significantly associated with collectivistic cultures, which typically discourage the experience and expression of HAN emotions (Kitayama et al, 2015).…”

“…We conducted two secondary analyses of data from two studies by Liew et al (2021aLiew et al ( , 2021b) using Spotify's energy (instead of danceability). Analysis 1 examined cultural differences in energy from top-50 charts in late-2018 (R 1 ) and mid-2019 (R 2 ), between East Asia (Japan, Taiwan, Hong Kong, and Singapore), and the Englishspeaking West (USA, UK, Australia, and Canada), for a total of N = 800 songs.…”

“…This was observed for rhythmic markers of musical arousal, namely danceability, but preliminary observations suggest that intensity markers of musical arousal, such as energy, may deviate from this trend. Liew et al (2021b) found that music consumed and produced in Japan, with lower affordances for HAN emotions, was much higher in energy than in the US, with higher affordances for HAN emotions. Similarly, Vidas et al (2021) found no significant differences in energy between top‐50 music from Pan‐Asian and Anglo‐Western countries.…”

Examining the Lyrical Content and Musical Features of a Crowd-Sourced, Australian Pandemic Playlist

“…Tempo is measured in beats per minute, while the other audio features are scored from 0.0 to 1.0 (https://developer.spotify.com/documentation/web-api/reference/#/operations/get-audio-features). Several recent studies have used the Spotify API variables to examine people's music listening behaviors (e.g., Liew et al, 2021; North & Krause, 2022; Panda et al, 2021; Vidas et al, 2021b).…”