Inside the Engine .
A small look at how the engine handles cultural data — what a record carries, the route from source to surface, and the gates that keep the bad stuff out. Enough to know what's holding it up. The deep internals stay deep.
From dataset to connection.
It's not a database — it's a graph. Every artist, every venue, every work gets a stable ID and is semantically linked. Mozart doesn't sit somewhere alone — he connects to The Magic Flute, the Marionette Theatre, and the Festspiele.
Four sides per record.
Every venue, festival, artist, work, production, and showtime carries the same four-sided structure. Edited independently, read together — so the API, the widget, and the signage all see one thing.
Source to surface.
Every record passes through the same five steps — same path, every time. A new country, language, or partner plugs into the same route without a rebuild.
One record per thing.
Five sources for the same venue collapse into one. Spelling variants, scripts, and abbreviations all match the same place — so what reads the data sees one truth, not five copies of it.
Data quality is measurable.
Every record gets a truth score from 0 to 100 — based on source plausibility, graph coherence, and cross-reference agreement. Public sector partners get the number they need to defend funding decisions.
Edit without breaking.
Your team edits one version, the network keeps another, and both stay current. You can rewrite a description for your audience without breaking what other partners see — and updates from the network keep flowing in underneath.