Fabiola C. Villanueva, Jonaz Juan C. Sayson, Christine D. Bandalan
In the digital streaming era, data-driven forecasting of music genre popularity has become increasingly important and significant in influencing both content creation and distribution. This study has investigated how audio features characterized by Spotify tracks are used by forecasting genre popularity trends and identifying which audio features are key drivers of listener engagement. After inspection, the dataset composed of 652 tracks from 2008-2024 containing high-popularity songs across Pop, Hip-hop, Latin, and Electronic genres. It also included nine audio features that were extracted and aggregated for temporal modeling which were namely, acousticness, danceability, energy, instrumentalness, liveness, loudness, speechiness, tempo, and valence. The dual-model system was developed using XGBoost for sonic characteristic analysis, a micro-level feature importance analysis at the track level. Whereas ARIMA handled music genre forecasting at a macro-level via walk-forward and time-aware validation. Key results include how XGBoost reveals genre-specific audio feature prioritization. Despite ∼70% of non-audio factors such as marketing and artist fame not considered for individual track popularity, XGBoost effectively isolates the remaining ∼30% from Spotify’s nine audio features to rank intra-genre success. Tracks were intentionally aggregated yearly into median genre popularity for ARIMA time-series forecasting, smoothing noise to reveal audio features for long term forecasting. This enables producers and beneficiaries to engineer genre-competitive tracks while strategically selecting and investing in stable or growing genres. This data-driven approach for a workflow elevates intuitive hit rates from ∼5% to 15–20% by providing production targets and 1–5-year trend horizon for informed navigation of the streaming music landscape. © 2026 Copyright held by the owner/author(s)
Department of Computer, Information, Sciences and Mathematics, University of San Carlos Cebu City, Cebu, Philippines
