Mapping Emotional Dynamics Through Music Consumption Patterns

Emotions significantly influence human behavior, impacting decisions, social interactions, and overall mental wellbeing. Understanding emotional dynamics over time is crucial not only for clinical applications, such as early detection of psychological distress, but also for enhancing user-centered services like recommendation systems and personalized advertising. Traditional emotion recognition methods rely on physiological sensors, facial expressions, or textual analysis—approaches that, while effective in controlled scenarios, suffer from issues of invasiveness, scalability, and privacy. In this paper, we propose a novel, non-invasive approach to infer users’ emotional flow over time by analyzing their music listening history. Music is deeply intertwined with emotions: individuals often select songs based on their current mood, and conversely, music can induce emotional changes. Leveraging this connection, we identify recurring listening habits using a spatio-temporal clustering technique that jointly considers musical similarity and listening time. Each habit is then associated with a corresponding emotion, enabling us to reconstruct a daily emotional profile over a typical week for each user. We validate our method through an experimental study involving 14 real users who provided their listening histories. Results show the system’s ability to capture individual emotional trends, highlighting its potential for both psychological monitoring and commercial applications in emotion-aware computing.