1. Official statistical layer
Use BPS and the Directorate General of Estate Crops for planted area, production, productivity, trade and time-series statistics.
2. Legal & origin layer
Use DGIP for Geographical Indication registration names and registration references. Do not infer legal GI status from marketing names.
3. Geospatial layer
Use BIG boundaries for province, regency/city, district or village polygons. Store geometry separately from coffee-origin records so it can be updated independently.
4. Agronomy layer
Use Ministry of Agriculture and research/extension sources for species, varieties, cultivation characteristics and technical references.
5. Sensory & lot layer
Treat tasting notes as lot-specific evidence whenever possible. Store cupping date, evaluator, sample preparation, crop year, producer and processing data alongside the sensory record.
6. Governance layer
Every record should carry source URL, source organization, retrieval/publication date, verification status, editor and last-updated timestamp.
Recommended production data schemaOrigin ID · origin name · province · regency · district · coordinates · boundary ID · species · cultivar · altitude range · process · harvest period · GI status · GI number · producer/cooperative · crop year · sensory record · source URL · source date · verification status.
Data governance
Use a layered source-of-truth model.
Coffee data is multi-dimensional. Statistics, legal GI status, geospatial boundaries, agronomy and sensory assessment should each come from the source best suited to that data type.