Data Sources

Where the Council's data comes from — and how it is controlled

Every published figure passes through a registered source, a set of automated quality checks and a controlled publication workflow before it appears on this site.

Collection

Registered data sources

The Council draws on national customs directorates, member country submissions, trade APIs and a dedicated price panel. Each source has a named owner, a reporting cadence and a documented coverage scope.

ITC Trade Map API

Statistics Unit · Monthly

pending

Bilateral trade, HS 0804.10 family

Last Updated
15 Dec 2025
Next Refresh
17 Jan 2026
Records
184,320
Type
api

Integration-ready: connector, field mapping and scheduling are configured; awaiting production API credentials.

Member Country Submissions

Membership Secretariat · Quarterly

healthy

Production, area, yield, prices

Last Updated
12 Jan 2026
Next Refresh
12 Apr 2026
Records
42,180
Type
file upload

CSV, Excel, JSON and XML uploads validated at ingestion before entering the review workflow.

National Customs Directorates

Trade Intelligence Unit · Monthly

healthy

Export and import declarations

Last Updated
9 Jan 2026
Next Refresh
8 Feb 2026
Records
96,540
Type
api

Records are reconciled against mirror statistics before publication.

Council Price Panel

Market Analysis Unit · Weekly

degraded

Wholesale and export unit prices

Last Updated
6 Jan 2026
Next Refresh
20 Jan 2026
Records
12,760
Type
survey

Two panel countries are behind schedule; index uses carried-forward values flagged in the metadata.

Contextual Agriculture Datasets

Research Unit · Annual

healthy

Climate, water, population and GDP context

Last Updated
28 Nov 2025
Next Refresh
11 Nov 2026
Records
8,940
Type
api

Used only for contextual overlays; never mixed with official council indicators.

Validation

Automated quality checks

Every payload — from an API pull or a member upload — is screened against seven checks before a human reviewer sees it. Blocking checks must pass before a submission can proceed; warnings are surfaced for reviewer judgement.

Schema conformity

blocking

All required columns present with the expected data types.

Value range

blocking

Values fall within plausible statistical ranges for the indicator.

Unit consistency

blocking

Units match the indicator definition (tonnes, hectares, USD).

Duplicate detection

warning

No duplicate country-year-indicator combinations in the payload.

Outlier screening

warning

Year-on-year change beyond three standard deviations is flagged for review.

Completeness

warning

Coverage of mandatory years and indicators for the reporting period.

Source attribution

blocking

Every record carries a named source and collection method.

Approval & publication

The controlled publication workflow

Data moves through five gated stages. Nothing reaches the public statistics platform without passing every stage in order.

  1. 1

    Ingested

    Data received from a source or member upload.

  2. 2

    Validated

    Automated quality checks executed and scored.

  3. 3

    Reviewed

    Checker has reviewed flags and confirmed accuracy.

  4. 4

    Approved

    Approved for release by the administrator.

  5. 5

    Published

    Visible on the public statistics platform.

In plain language

Collection: figures start life either as an automated feed from a registered source (customs, trade APIs, contextual datasets) or as a file uploaded by a member country's statistical focal point.

Validation: seven automated checks screen every payload for schema, range, unit, duplicate, outlier, completeness and source-attribution issues before any reviewer sees it.

Approval: a named Checker reviews flagged rows, and a Council Administrator gives final sign-off before release.

Publication: only approved, published records feed the public statistics platform, indices and downloadable datasets.