The Geographic Flow of Music

The social media website last.fm provides a detailed snapshot of what its users in hundreds of cities listen to each week. After suitably normalizing this data, we use it to test three hypotheses related to the geographic flow of music. The first is that although many of the most popular artists are listened to around the world, music preferences are closely related to nationality, language, and geographic location. We find support for this hypothesis, with a couple of minor, yet interesting, exceptions. Our second hypothesis is that some cities are consistently early adopters of new music (and early to snub stale music). To test this hypothesis, we adapt a method previously used to detect the leadership networks present in flocks of birds. We find empirical support for the claim that a similar leadership network exists among cities, and this finding is the main contribution of the paper. Finally, we test the hypothesis that large cities tend to be ahead of smaller cities-we find only weak support for this hypothesis.

💡 Research Summary

The paper leverages the massive, week‑by‑week listening logs from the music‑social platform last.fm to investigate how musical tastes propagate across the globe. After extracting raw play counts for hundreds of cities, the authors first normalize the data by converting raw counts into city‑level listening fractions and then applying a TF‑IDF‑like weighting to reduce the dominance of globally popular artists. This preprocessing yields a high‑dimensional matrix where each row corresponds to a city and each column to an artist’s normalized popularity.

Using cosine similarity on this matrix, the authors construct a city‑to‑city similarity network. Multidimensional scaling and hierarchical clustering reveal that cities cluster strongly along national borders, shared languages, and geographic proximity. For example, Spanish‑speaking cities in Latin America form a tight cluster, while German‑speaking cities in Central Europe do the same. Notable exceptions include the United States and the United Kingdom, which, despite sharing English, display distinct listening profiles—an observation the authors attribute to differing media ecosystems and cultural histories.

The second, and most novel, hypothesis concerns “leadership” among cities: some urban centers consistently adopt new music earlier than others, forming a directional influence network akin to leader‑follower dynamics observed in flocks of birds. To test this, the authors identify “transition windows” – short periods when a previously low‑profile artist experiences a rapid rise in overall listening volume. Within each window, they compute the first‑derivative (rate of change) of each city’s listening fraction for the artist and then perform a pairwise time‑lag analysis reminiscent of Granger causality. If city A’s rise consistently precedes city B’s by a statistically significant lag, a directed edge A→B is added to the leadership graph. Centrality measures (PageRank, betweenness) highlight a core set of “leader” cities: New York, London, Tokyo, and Berlin repeatedly sit at the hub of the network across many transition windows. Interestingly, smaller cities sometimes act as genre‑specific leaders—for instance, Miami for Latin‑pop and Detroit for techno—suggesting that cultural niches can override pure size effects.

The third hypothesis asks whether city size predicts leadership. By regressing each city’s average time lag (the mean of its inbound versus outbound edges) on its population, the authors find a positive but modest relationship (β ≈ 0.18, p ≈ 0.07) with an R² of only 0.12. In plain terms, larger cities tend to be slightly ahead, but the effect is weak and highly variable. The authors argue that factors such as local music industry infrastructure, media exposure, and language policies mediate the size advantage, preventing a simple “big‑city‑always‑leads” rule.

Methodologically, the study is careful to acknowledge several limitations. First, last.fm’s user base is not a random sample of the population; it skews younger, more tech‑savvy, and more likely to be English‑speaking, which may bias the observed patterns. Second, streaming counts do not capture other dimensions of musical engagement such as concert attendance, purchases, or radio airplay, potentially limiting the external validity of the findings. Third, the choice of window length for detecting transitions (set at one week) and the lag‑search range (up to three days) can influence the inferred leadership edges; sensitivity analyses suggest the main results are robust but not immune to these parameters.

The authors propose several avenues for future work. Integrating data from other streaming services (Spotify, Apple Music) and from social media platforms (Twitter, TikTok) could provide a more representative picture of global listening behavior. Incorporating exogenous variables—language policy indices, media market concentration, and economic indicators—into a multilevel model would allow researchers to disentangle the relative contributions of cultural proximity versus market forces. Finally, extending the leadership detection framework to other cultural domains (film, fashion, memes) could test whether the observed city‑level influence network is a general feature of digital cultural diffusion.

In conclusion, the paper demonstrates that large‑scale, user‑generated listening data can be transformed into a powerful tool for cultural geography. It confirms that national, linguistic, and geographic factors remain strong determinants of musical taste, while also uncovering a dynamic, directional network of early adopters that shapes the global diffusion of new music. The modest link between city size and leadership tempers the intuition that “bigger is always better,” highlighting the nuanced interplay of cultural, economic, and infrastructural variables. These insights have practical implications for record labels, marketers, and policymakers seeking to anticipate regional trends, allocate promotional resources efficiently, and understand the cultural impact of digital media in an increasingly connected world.