About Rock Atlas

The mission

Rock Atlas is one part encyclopedia, one part atlas. Every band and solo artist has a quick-facts page or a long-form deep-dive article — that's the encyclopedia side. The five lenses (charts, map, eras, genres, influence) let you explore the same data five different ways — that's the atlas side.

Everything here is derived from public sources we can name and re-fetch: Wikipedia (article summaries and pageview velocity), Wikidata (formation/dissolution dates, genre taxonomy, the P737 influence graph, member rosters, country of origin), MusicBrainz (canonical discographies and platform IDs), Last.fm (cumulative listenership), and Wikimedia Commons (photography, where freely licensed). No data is invented; everything you see can be traced back to a public record.

How the rankings work

Three charts, three angles on the same question — who shaped rock?

• Bands — five hundred groups and duos, hand-tuned from decades of music writing and weighted toward influence over commercial reach. Each entry has a long-form article.
• Bands & Solo Artists, together — bands and solo artists in the same chart, ordered by a weighted score: 30% Last.fm listeners + 15% playcount + 10% Wikipedia pageviews + 30% PageRank on the Wikidata influence graph + 15% career longevity. Pre-1980 formations get a small canon-favoring multiplier to counteract the streaming-era recency bias.
• Solo Artists — same scoring, filtered to solo names only. The singer-songwriters, frontpeople, and instrumentalists who shaped rock outside band contexts.

The two algorithmic lists are produced by a reproducible pipeline. They're validated against a hand-checked canonical set — Beatles in the top 5, Led Zeppelin in the top 15, etc. — that hard-fails the run if the algorithm wanders. The full methodology, including data sources, the scoring formula, the weights and why we chose them, and the algorithm's known biases, is on the methodology page.

What counts as a rock artist

We include rock and its subgenres — from 1950s rockabilly through contemporary post-rock, math rock, sludge metal, and shoegaze — and exclude pure pop, hip-hop, country, and electronica unless an artist is unambiguously classed as rock.

For algorithmic discovery, an artist enters the candidate universe if their Wikidata genre list places at least 25% of their genres in the rock subtree (or 50% if pop / R&B / hip-hop markers are present). A small list of canonical bypasses (Prince, Steely Dan) covers artists whose Wikidata is dominated by adjacent genres but who are unambiguous rock canon.

Editorial voice

Articles are written like documentary narration: informative, neutral, historically grounded. No clickbait, no opinion-heavy editorializing, no shallow summaries. Each band entry covers formation, breakthrough, peak era, musical style, major albums, signature songs, influence, legacy, and lesser-known facts.

Long-form articles for the curated 1000+ bands and artists have rolled out in batches

What's still missing

• Critic-list data — Rolling Stone 500, Pitchfork, NME, Acclaimed Music. Their data is paywalled or behind aggressive scraping defenses; we approximate via Last.fm + pageviews.
• The influence visualization — the data is built; the force-directed graph rendering is in progress.
• Cover-art licensing — discography sections show MusicBrainz metadata but not album art beyond the iTunes preview thumbnails.

The stack

The site is a static build — no runtime database, no per-request rendering. Pages are generated by Astro at build time from typed TypeScript data files; images are processed locally with Sharp. The ranking pipeline runs as a sequence of TypeScript scripts that fetch from Last.fm, Wikipedia, Wikidata, and MusicBrainz, cache the results to JSON, and emit the final master lists.