Case study
Record Collection Analytics
A public, sanitized view of what's actually in my record collection: decade, genre, and format breakdowns, with no private inventory or prices exposed.
What it is
This page reads a sanitized snapshot of my Discogs collection and turns it into something worth looking at instead of a private inventory list. The current snapshot has 216 active records, 200 LPs and 8 singles, and the decade breakdown is not close: the 1970s and 1980s carry 62 and 61 records between them, more than half the shelf, with the 1990s as a real trough at just 4. Rock is 167 of the 216. Jazz is a distant second at 35, which is a higher share than I would have guessed before I actually counted.
Why I built it
Shelf management alone did not need a website. What I actually wanted was the shape of the buying, past the raw count: which decades I keep gravitating back to, where a genre I think of as central turns out to be thin, and where the catalog is honestly a little messy. A public version of that has to keep the useful signal while dropping anything that reads as a private inventory, so no prices, and no exact addition dates.
What it shows
Chart views by decade, genre, and format, built off the same
discogs-collection.json snapshot the rest of the site’s music pages pull
from. The snapshot also surfaces a small rotating set of featured records, at
the moment things like At Fillmore East, Damn the Torpedoes, and
Synchronicity, which is a fair cross-section of what actually fills the
1970s and 1980s numbers above.
Technical shape
The public site reads a static snapshot rather than live Discogs inventory,
which keeps it fast, keeps it on static hosting, and keeps it safe to
publish. The actual sync happens in Discogs Spinner, one dplayer export
run producing a JSON file that a small build script turns into the decade
and genre breakdowns above; this page never talks to Discogs directly and
never sees anything that export did not already choose to include.
What I learned
The useful question was never how much of the collection to show. It was which views actually tell me something I did not already know, versus which ones just turn the page into an inventory dump with extra charts.
What’s next
The snapshot behind this page is from 2026-05-27 and has not been refreshed since, so the numbers above are already a few months stale relative to whatever is actually on the shelf now. Regenerating it, and adding a visible “as of” date on the page itself so that staleness is not something a reader has to guess at, are the next real improvements.