Methodology
Data methodology
AgricultureID separates stable, descriptive facts from volatile statistics, and never presents a number without its year, unit, source, and dataset version.
Sources
Country statistics are drawn from World Bank Open Data (World Development Indicators), which is openly licensed (CC BY 4.0) and compiled from FAO and national statistical sources. We prefer an open, versioned, machine-readable source over scraping dashboards or search-engine results.
Immutable snapshots
Each indicator is captured as a dataset snapshot: a stored copy recording the dataset version (the provider’s last-updated date), the retrieval date, coverage, and an integrity checksum. A changed dataset version is saved as a new snapshot — historical snapshots are never overwritten.
How values are presented
- Every value carries its own year and unit.
- Values are described as dated, revisable statistics, not evergreen prose. We do not hard-code a specific figure into descriptive text.
- Modeled indicators (for example, employment in agriculture) are flagged as estimates, not observed counts.
- We do not rank countries using values from incompatible years, units, or methods, and we flag indicators (such as index numbers) that are not comparable across countries as absolutes.
Derived values
A value computed from other statistics (for example, arable land as a share of agricultural land) is labelled as derived and records its formula, formula version, inputs, and the source versions of those inputs. Derived values are never presented as observed measurements.
Limitations
Definitions, coverage, and revision cycles differ between datasets, so figures are not always comparable across sources or years. Country coverage is partial and some series are sparse. For authoritative, current figures, consult the primary dataset linked on each page.