← Back to Work

// 25 / OPEN DATA  ·  02 Nov 2026

First Pressing — a Career
as a Shape

TypeBrowser tool · live
StackSingle HTML file · vanilla JS
DataiTunes Search API (free, no key)
BackendNone

// the idea

A discography is a list. A career is a shape.

Every music site will give you an artist's albums as a list, newest first, with a star rating. None of them will show you the thing a list hides: the time between. Four albums in five years and then nothing for six. A debut, a quick second, and a decade before the third. The run of records that came out so fast they blur. The silences are half the story, and a list throws them away.

First Pressing draws the time. Type an artist and every album goes on a horizontal line by release date — studio albums bright and full-size, live records and compilations dimmer — with the gaps between studio albums drawn and labelled. Under it, the albums group themselves into eras, each with a one-line read. Click any sleeve and its tracks appear as bars by length, which is enough to see a 1970s two-sides shape, a 1990s 70-minute one, or a streaming-era run of two-minute tracks, without being told. Add a second artist and the two careers stack on the same years.

A concrete example

Pick any long career and the timeline tells you something the list didn't. The pattern that shows up again and again:

first pressinga typical long career
1979 ─●─●──●──●        four albums in five years — the first era
      └── 4 yrs 11 mo ──┘
1989 ────────────●     one record, then
                  └──── 10 yrs 1 mo ────┘  the long silence
2004 ──────────────────────────────●   the late return

eras     1979–84 · 4 albums, one every 1.9 yrs, 10 tracks on average
         1989 · 1 album, 12 tracks
         2004 · 1 album, 14 tracks
calls    longest silence 10 yrs 1 mo · quickest follow-up 1 yr 2 mo

Read that against what you know of the artist and the gaps have names: the label dispute, the solo years, the illness, the comeback. The tool doesn't know any of that and doesn't pretend to. It draws the shape; you bring the story.

Try it — no sign-up

Hit one of the names under the box, or type your own. Then + Compare with… and add a contemporary to see two careers on the same years.

Open First Pressing →

// what it does

Four panels

01

Artist

Type-ahead suggestions from the catalogue. Three filters: show or hide live albums, compilations and EPs; hide reissues and deluxe editions (on by default, because a popular catalogue has four versions of every record); include singles. At the top of the results, five numbers: studio album count, career span with the average spacing, longest silence, quickest follow-up, tracks per album.

02

The timeline

One line, scaled to the career, sleeves placed by release date. Gaps of eighteen months or more between studio albums are drawn as dashed brackets with the length written in; four years or more turns red. Sleeves that would overlap get staggered stems. A second artist goes below the line in purple. Share link opens on the same pair; save-as-image renders a clean chart without the artwork (which can't be drawn to a canvas cross-origin).

03

The eras

Studio albums cut into runs wherever there's a gap of three years or more. Each era: the years, how many albums, how often, the average track count, and the sleeves. Beneath, the call-outs: most productive stretch, longest and shortest album, first and latest, and the side-by-side summary when there are two artists.

04

The shape of an album

Click a sleeve: tracks as bars by length, the opener and closer in gold, the longest in purple, total running time and a one-sentence read of the shape — vinyl-era, CD-era, or streaming-era.

// how it was built

Build notes

The iTunes Search API is the quiet hero of free music data

No key, no account, no rate-limit paperwork, and it covers nearly everything with a release date. Three calls do the whole job: a search for the artist, a lookup for all their collections, and a lookup per album for tracks. Artwork comes back as a URL you can resize by editing the path. It's served as JSONP, which is why the page loads it with a script tag rather than fetch — an old technique, but exactly right here.

De-duplication is most of the work

A big catalogue is a mess of editions: the album, the remaster, the deluxe, the 40th anniversary, the mono, the stereo, the one with the bonus disc. Left alone, a 1970s band looks like they released forty records. The page strips parenthetical qualifiers to get a base title, drops anything whose title says deluxe, remastered, anniversary and so on when a plain version exists, and collapses what's left to the earliest date per title. A separate list of words — live, in concert, best of, anthology, sessions — marks a release as not-a-studio-album so it's drawn small and excluded from the gaps. It's heuristics, it's visible in the source, and the filters let you switch it off when it guesses wrong.

firstpressing.htmljavascript
// keep a reissue only if there's no plain version of the same title
list = list.filter(c => !c.isReissue || !list.some(o => o !== c && o.base === c.base && !o.isReissue));
// then collapse same-title duplicates to the earliest release
const seen = new Map();
for (const c of list) if (!seen.has(c.base) || c.date < seen.get(c.base).date) seen.set(c.base, c);

Eras are a single threshold

A new era begins after a gap of three years or more between studio albums. That's it — one number, chosen because it separates "the next record" from "a return." It's wrong for some artists and right for most, and it's the first thing in the source if you want a different one.

The dates are the catalogue's, not history's

The store's release date is sometimes the date the record went up digitally rather than the original pressing, especially for older catalogue that arrived in a batch. For most artists it's right; for a few the 1960s look like 2005. The page says so at the bottom rather than silently trusting it, and the earliest-per-title collapse helps when both dates are present.

Lines
~330 incl. CSS
API calls
2 per artist + 1 per album clicked
Key
None
Stored
Nothing

// what this demonstrates

Why this matters

It's a small argument about data that already exists. The catalogue has been sitting behind a free, keyless API for fifteen years; every music app uses it to sell you the record. Nobody used it to show you the time. The gap between "the data is available" and "someone drew the obvious picture" is where a lot of useful tools live, and this is one.

It also belongs with Weather Window in the toolkit's "for everyone" group: no AI, a free public source, one visual idea done properly, and rules you can read in the source and disagree with.

Where it could go next

Running time over the years

Plot album length and track count by year across the whole career — the CD bloat and the streaming shrink as a line.

Label and producer

MusicBrainz has both, free and keyless; the eras would gain names.

Band members

Also MusicBrainz: who joined and left, drawn under the timeline, so a line-up change sits next to the gap it caused.

Four at once

A scene view — the four big acts of a city or a decade, stacked, to see who was answering whom.

Source

The tool is the source — view the page and it's all there, MIT. firstpressing.html