Methodology

Data & methodology

In general terms: this map is built from artist-relatedness data, which artists' fans also follow which other artists, in the spirit of Spotify's own "Fans Also Like" framing. This page describes the shape of the pipeline, not the collection mechanics.

5,322Artists in this snapshot
22Regions
92Scenes at default granularity
2026-08-13Snapshot date
MonthlyRefresh cadence
LeidenCommunity-detection algorithm

How the map is built

1. Seed pool
A fixed pool of seed artists anchors the graph.
2. One-hop expansion
Each seed's "fans also like" neighbors are added, one hop only; the graph does not expand recursively from newly-discovered artists.
3. Two graph views
An undirected view (who shares neighbors with whom) drives community detection; a directed view (who points at whom) drives the "magnet"/gravity read.
4. Community detection
Leiden clustering, fixed random seed for reproducibility, tuned so scenes land at a roughly scene-sized granularity rather than a handful of giant blobs or hundreds of singletons.
5. Two anchors per region/scene
The core / magnet (most affinity-central member) and the best-known member (highest Spotify monthly-listener count among members who themselves clear a within-region/scene relevance floor, not a bare fame ranking) are reported side by side, since they can diverge.
6. Layout
A force-directed layout positions artists so affinity pulls related artists together; positions are fixed per snapshot.
Not affiliated with Spotify. This is an independent project. It is not produced, endorsed, or reviewed by Spotify.

Refresh cadence

The map is refreshed monthly. Affinity relationships move slowly compared to, say, weekly chart positions, so monthly keeps the map current without churning scene membership or naming more than the underlying reality does.

Because the graph itself can drift between snapshots (an artist can move scenes, a scene's best-known member can change), the exact scenes, regions, and best-known names shown on this site's example pages reflect the 2026-08-13 snapshot and may shift after the next refresh.

What's next

A monthly cadence opens the door to a future month-to-month time-lapse view of how scenes evolve, not built yet and not a promised feature. The real ramifications are mainly keeping a scene's identity stable across snapshots so it reads as evolution, not noise.

Placeholder: exact data-retention and correction-handling policy beyond contacting us isn't decided yet

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.