Sonograph
Tracks · Artists · Genres
An explorable network of 581 tracks linked to the 351 artists who performed them and the 229 genres they belong to — 6,369 links in total. CineGraph's musical twin: same visual language, same interaction patterns, a different dataset.
Tracks Artists Genres Albums Producers Labels
Genres act as high-degree attractors that hold the layout together; artists and producers form the visible clusters. Collaborations and covers let one artist bridge genre neighbourhoods. Drag to pan, scroll to zoom, click a node for its details and to highlight its connections, or use the search box to jump straight to one.
Every node is also listed under /notes — tracks, artists, genres, albums, producers and labels — each name opening that node on the interactive graph.
The sample
Recording metadata comes from MusicBrainz, with optional Discogs, Genius, Last.fm and Wikidata enrichment, from the seed list in config.yaml.
Most-connected genres
| Genre | Tracks |
|---|---|
| Pop | 163 |
| Rock | 156 |
| Indie | 140 |
| Folk | 127 |
| Country | 105 |
Most-represented artists
| Artist | Tracks |
|---|---|
| Taylor Swift | 16 |
| Ariana Grande | 10 |
| Linkin Park | 10 |
| Paramore | 9 |
| Arctic Monkeys | 8 |
Which genres travel together
The bipartite track–genre graph projected onto genres alone: 229 genres, 1821 weighted pairs, where a weight is the number of sampled tracks carrying both.
| Genre pair | Shared tracks |
|---|---|
| Alternative Rock + Rock | 59 |
| Pop + Rock | 44 |
| Alternative + Rock | 39 |
| Indie + Indie Rock | 38 |
| Indie + Rock | 33 |
How it was built
A Python pipeline fetches recording metadata (MusicBrainz as the canonical
backbone, with Discogs, Genius, Last.fm and Wikidata as enrichers), caches
every raw API response to disk so builds are reproducible, models the result
as a networkx.DiGraph of six node types, then renders it as the
interactive graph above. A graph.json dump of the same network
is written alongside it.