A demo by Beatriz Bustos · Knowledge / Ontology Engineering
Español · Try the visual walkthrough
A laboratory buys gloves from three suppliers. Its report groups everything as “gloves”, but the team needs to distinguish nitrile and latex.
A new catalog introduces those two categories. Six product links now need updating without losing or misclassifying their 180 purchase lines.
- Before: a diagram connects suppliers, the old category and the purchase report.
- Change: the app traces six affected links to their 180 dependent lines.
- Decide: read the supplier’s material and choose nitrile or latex. Incompatible choices are explained and not applied.
- Result: the report separates 90 nitrile lines and 90 latex lines. No purchase or amount is lost.
The useful outcome: review affected product links once, update all dependent purchases and retain the reason for each decision. No measured customer time saving is claimed.
How can a technical team verify it?
Open “Technical evidence” on the result screen:
- Inspect the six original SKUs, supplier material and review decisions.
- Execute SQLite in the browser. Query the same 180 lines and verify the same report and amounts.
- Inspect the same decisions in RDF/SKOS, OWL/RDF/XML and Cypher. Cypher is a preview; no Neo4j server is deployed.
This case is an explicitly scoped part of an 8,640-line synthetic regression dataset. The demo does not claim to have resolved all those other catalog changes.
Data and validation · Code map
npm ci --prefix 04-app
npm run dev --prefix 04-app
npm test --prefix 04-appSuppliers and purchases are fictional. Reviews last for the session. No paid APIs are needed. Importing company data and sharing team reviews are not implemented yet.