How has popular music evolved time and space? Have modern technology and the internet fundamentally changed the way we consume and share music? Until recently, these questions remained largely unanswered due to limited data and analytical tools. The digital age and advances in machine learning have changed everything, opening up exciting new avenues for scientific research on music.
Throughout my PhD, I've assembled an extensive dataset comprising some 250,000 popular songs from around the globe, spanning from the 1950s to the present day. This collection is complemented by a massive ongoing radio monitoring project covering 220 countries, which has already accumulated tens of terabytes of audio data of spoken conversations and music. Applying cutting-edge machine learning tools to this vast library, including music information retrieval, natural language processing, and social network analysis, my collaborators and I sought to characterise the global history of popular music through a data-driven approach.
*GlobalPOP website launch and data open release coming in 2025*
Lee, H., Çelen, E., Harrison, P. M. C., Anglada-Tort, M., van Rijn, P., Park, M., Schönwiesner, M., & Jacoby, N. (2025). GlobalMood: A cross-cultural benchmark for music emotion recognition. Proceedings of the International Society for Music Information Retrieval Conference (ISMIR 2025). https://doi.org/10.48550/arXiv.2505.09539 [PDF]
Anglada-Tort, M., Lee, H., Krause, A. E., & North, A. C. (2023). Here comes the sun: Music features of popular songs reflect prevailing weather conditions. Royal Society Open Science, 10(5), 221443. https://doi.org/10.1098/rsos.221443 [PDF]
Lee, H., Hoeger, F., Schoenwiesner, M., Park, M., & Jacoby, N. (2021). Cross-cultural mood perception in pop songs and its alignment with mood detection algorithms. International Society for Music Information Retrieval Conference (ISMIR 2021), 366–373. https://doi.org/10.5281/zenodo.5625680 [PDF]
Presented at CES 2022: "GlobalPOP Studying the evolution of pop music across 55 countries since the 1950s"
From gossips about your peers to viral TikToks that spread across society, patterns of contagion follow similar rules. As such, by knowing which musical trends catch on 'where' and 'when', we can make probabilistic estimates about the possible musical connections between places, and ultimately predict the patterns of spread.
Thanks to groundbreaking new tools like PsyNet, we can also now conceptualise these scenarios in an experimental setting. Picture an online version of the Chinese whispers game, where we can track exactly how messages change as they pass along between people. We can even create miniature virtual societies to observe how manipulating the breadth of interactions between people might shape group behaviour. By combining these clever experiments with real-world data, we're beginning to understand how ideas and cultural trends catch on and spread.
Lee, H.†, Anglada-Tort, M.†, Sobchuk, O., Rijn, P. van, Schönwiesner, M., Tchernichovski, O.*, Park, M.*, & Jacoby, N.* (2024). Rapid reshaping of cultural boundaries during war across 1,423 cities. Preprint https://doi.org/10.31234/osf.io/7b98u [PDF]
Anglada-Tort, M., Harrison, P. M. C, Lee, H., & Jacoby, N. (2023). Large-scale iterated singing experiments reveal oral transmission mechanisms underlying music evolution. Current Biology, 0(0). https://doi.org/10.1016/j.cub.2023.02.070 [PDF]
Marjieh, R.†, Rijn, P. V.†, Sucholutsky, I.†, Sumers, T., Lee, H., Griffiths, T. L.*, & Jacoby, N.* (2023). Words are all you need? Language as an approximation for human similarity judgments. International Conference on Learning Representations (ICLR 2023). https://openreview.net/forum?id=O-G91-4cMdv [PDF]
van Rijn, P.†, Lee, H.†, & Jacoby, N. (2022). Bridging the prosody GAP: Genetic Algorithm with People to efficiently sample emotional prosody. Proceedings of the Annual Meeting of the Cognitive Science Society (CogSci 2022). https://escholarship.org/uc/item/59q714w6 [PDF]
Why do some songs just click with us while others don't? Or why certain paintings are more appealing to look stare at? When you and another person share the same taste in music, do you also tend to like the same films or clothes? Truth is, we're still trying to figure out exactly how our tastes work, and whether liking one thing means we'll like something similar in a different area.
Through analysing millions of behavioural patterns in the real world and conducting cross-cultural surveys about preferences across multiple domains, I seek to understand how our environment and experiences shape our cultural tastes. For instance, during my research internship at Deezer, I analysed the listening patterns of over 2.5 million users across France, Brazil, and Germany, revealing how a city's demographic composition and size influence the diversity of its residents' musical tastes.
Lee, H., Jacoby, N., Hennequin, R.*, & Moussallam, M.* (2025). Mechanisms of cultural diversity in urban populations. Nature Communications, 16(1), 5192. https://doi.org/10.1038/s41467-025-60538-2 [PDF]
Lee, H., Van Geert, E., Çelen, E., Marjieh, R., van Rijn, P., Park, M., & Jacoby, N. (2025). Visual and Auditory Aesthetic Preferences Across Cultures. Proceedings of the Annual Meeting of the Cognitive Science Society (CogSci). https://doi.org/10.48550/arXiv.2502.14439 [PDF]
Çelen, E., van Rijn, P., Lee, H.*, & Jacoby, N.* (2025). Are Expressions for Music Emotions the Same Across Cultures? Proceedings of the Annual Meeting of the Cognitive Science Society (CogSci). https://doi.org/10.48550/arXiv.2502.08744 [PDF]
Marjieh, R.†, Harrison, P. M. C.†, Lee, H., Deligiannaki, F., Jacoby, N. (2024). Timbral effects on consonance disentangle psychoacoustic mechanisms and suggest perceptual origins for musical scales. Nature Communications, 15(1), 1482. https://doi.org/10.1038/s41467-024-45812-z [PDF]
Lee, H., Hennequin, H, & Moussallam, M. (2023). Understanding individual and collective diversity of cultural consumption through large-scale of music listening events. Computational Humanities Research Conference (CHR 2023). https://ceur-ws.org/Vol-3558/paper3539.pdf [PDF]