Article
Comprehensive Analysis of Music Streaming Behaviour Using Spotify data and Insights
Music streaming platforms have transformed listening from an album-oriented activity into a continuously personalized, data-driven experience. Spotify generates a rich behavioural footprint involving listening histories, track characteristics, artist preferences, playlist interactions, temporal patterns, discovery activity, and contextual signals. However, existing research frequently treats music consumption as separate problems such as recommendation, popularity prediction, user profiling, or audio-content analysis. This review synthesizes recent research on music-streaming behaviour with particular emphasis on Spotify data and identifies the limitations of isolated analytical approaches. Six interrelated dimensions are examined: multi-source music-behaviour representation; multi-temporal user preference modelling; content-driven and collaborative recommendation; contextual, social, and new-release discovery; explainability, diversity, and popularity bias; and adaptive learning for changing user preferences. Recent Spotify research demonstrates that generalized user representations can combine multimodal signals across different temporal scales and support multiple downstream personalization tasks, while studies of new-release and socially motivated listening demonstrate that recommendation relevance depends on content age, community, culture, and timing. Based on this synthesis, this review develops a Unified Spotify Behaviour Analytics Framework (USBAF) consisting of data acquisition, behavioural representation, intelligence, adaptive explanation, and insight layers. The review concludes with open challenges and a staged roadmap toward context-aware, explainable, diverse, continually adaptive, and generative music-streaming intelligence.
Full Text Attachment
