Data repository

Data with provenance, revision history and a reason to trust every field.

The Lab treats a dataset as a versioned research object. Coverage, definitions, source perimeter, transformations, codebook, validation gap, corrections and revision history travel with the file.

Search the Evidence Terminal →Machine-readable release registry →Statistical vintages →

Live prospective register

MENA Statistical Vintage Register

First live record

MSVR-0001 · Saudi Arabia Q2 2026 real GDP.

Flash headline

−4.8% year on year.

Frozen

20 August 2026.

State

LIVE · later comparable vintage pending.

The register preserves what an official statistic said at a specific point in time. Later values are appended rather than used to overwrite the earlier information set. MSVR-0001 freezes GASTAT's Q2 2026 flash estimate before a later comparable vintage has been attached to the Lab record.

Read MSVR-0001 Register JSON Record CSV Protocol v0.1

Claim boundary

A LIVE record is not a forecast that a statistic will be revised. The later value may rise, fall or remain unchanged. Revision magnitudes are calculated only after a directly comparable later vintage is observed.

Release candidate

MENA Economic Narrative & Market Stress Observatory / Official Price Release Observatory

Coverage

48 geography-month release-candidate rows.

Observations

195 numeric observations.

Sources

30 official release links.

State

v0.3.0-rc1.

Current infrastructure preserves source-release linkage and revision-aware evidence state. Independent release-specific validation is not inferred from institution-level reproduction totals.

Data / code Project Validation Jordan evidence brief

Known limitations

The current state is explicitly a release candidate. A frozen release and publishable release-specific external-validation mapping remain separate gates.

Live public preview

MENA Funding & Procurement Intelligence

Records

11,799 normalized award records.

Perimeter

Current headline count uses World Bank structured sources.

Released

13 August 2026.

Evidence

Official source · real public data.

Manifest Schema Source register QA Code

Known limitations

UNGM, EU and GIZ are not included in the current headline count. The preview is not a census of all MENA procurement.

Audited collection / provenance alpha

MENA Firm AI Adoption and Labor Adjustment

Firms with numeric employment

38 / 50 firms.

Collected

176 / 400 numeric employment firm-years.

Provenance

168 / 176 employment observations with SHA-256 source bindings.

State

v0.2.0a2.

Coding schema Data / code Project Audit state

Known limitations

The empirical panel is not frozen for causal estimation. Collection remains incomplete and the independent second-human coding gate is pending; preferred causal estimates are not treated as released findings.

Infrastructure roadmap

These are dataset classes in the research programme roadmap, not falsely claimed released datasets.

Statistical Release Calendar

When does official MENA data actually arrive?

Institution · indicator · frequency · typical lag · source URL · historical archive availability · revision policy.

Labour Indicator Harmonization

Definitions preserved, not flattened.

Country × year × source × definition × sex × age × indicator, with national/harmonized distinctions retained.

Data Revision Atlas

First vintage versus revised vintage.

MSVR now supplies the prospective time-of-knowledge layer; the Atlas measures comparable revision chains and preserves methodological breaks.

Institutional Source Registry

Map the official-data system itself.

Central banks · statistical offices · ministries · regulators · budget/open-data portals · archives · machine readability · language · update frequency.

Required release fields

title · version · release_date · coverage · variables/schema · publisher/source perimeter · license when issued · download · codebook · methodology · source register · code · QA · limitations · validation state · citation · corrections · revision history

Institutional integrity

The Lab reports 2–3 public datasets at institution level, 50 dataset downloads, 100% source/provenance coverage across published datasets and 100% code/versioning coverage across reproducible outputs. Those totals do not erase release-specific states.

Institutional metrics →