Cadastre
Cadastre

Methodology

Scope

Cadastre catalogues publicly available data on land tenure, property markets, cadastral systems, and the built environment across African countries. Country is a top-level routing dimension; the schema is continental, not country-shaped.

Provenance

Every dataset has a registry entry stating publisher, license, fetch method, fetch path, processed path, update frequency, and last-fetched timestamp. Every chart cites its dataset; every entity page provides BibTeX, APA, and Chicago citations.

Languages

French is canonical for legal corpus. English and French are otherwise both first-class. Pages are sourced separately per language; translations are not regenerated by machine.

Content schemas

Country, City, Quarter, Legal concept, Glossary term, and Dataset entities each have a strict YAML frontmatter schema validated at build time. A build that contains malformed frontmatter fails.

Data pipeline

Raw source files live in data/raw/ and are never modified. ETL scripts in data/scripts/ are idempotent, respect robots.txt, and emit a manifest.json next to each output. Builds fail if a "live" dataset is older than its freshness window.

Limits

Cadastre does not invent figures. Where a dataset is pending ingestion, the corresponding chart renders a placeholder noting that data is not yet available. The transaction registry contributed by partners is anonymised before any aggregation reaches a public surface.