How ONYX is built

Every number shows its receipts.

ONYX turns 70+ canonical sources into decision-ready intelligence on the Nigerian economy. The point of this page is simple: you should be able to defend any ONYX figure to an investment committee. Here is exactly how each one is made — and how our composite indices are statistically validated, not just asserted.

The ONYX data cube: source, geography and timeThree planes meeting at right angles, each a four-by-four grid. One axis is the publishing source, one is geography from state down to local government area, and one is time. An observation in ONYX is addressed by all three together.SOURCETIMEGEOGRAPHY

Three guarantees on every figure

Triangulation

Every figure is cross-checked across at least two canonical sources — NBS, CBN, DMO, NUPRC, IMF, the World Bank. Where they diverge, the divergence is flagged, never silently averaged.

Provenance

Every observation carries its source, report, publication date, and access date. Each value is traceable to the original canonical reference — visible in the terminal, not buried.

Reconciliation

When sources disagree, the canonical-source value is published and the alternates are kept as superseded receipts. The resolution is auditable; there are no hidden choices.

How the composite indices work

ONYX publishes 0–100 composite indices (Investment Climate, Fiscal Stress, FX Pressure). Each component is scored by its causal percentile rank against its own history up to that point — an expanding window, so a year is never scored using data that did not yet exist (no look-ahead bias). Trending nominal levels (e.g. the naira rate) enter as year-on-year change so they cannot saturate the score. Components are then weighted and summed; the weights are deliberate analytical priors, documented and open to challenge.

Validated, not asserted

We run the robustness checks the OECD/JRC Handbook on Composite Indicators treats as essential. The verdict: the indices hold up.

Investment Climate
45.8/ 100
90% weight-uncertainty band: 34.957.0 · higher = better
Fiscal Stress
63.5/ 100
90% weight-uncertainty band: 61.066.1 · higher = more stress
FX Pressure
14.2/ 100
90% weight-uncertainty band: 10.917.3 · higher = more pressure
Weight uncertainty (Monte-Carlo)

We resample the weights 5,000× around their priors and recompute. The scores move only modestly — the published reading does not hinge on any one weighting choice. The band on each gauge above is that 5th–95th-percentile range.

No double-counting (correlation + PCA)

Across every index, no two components correlate above 0.70 — they measure distinct things, so nothing is silently counted twice.

Robust to the scoring method

Swapping percentile-rank for a z-score normalisation barely moves the rankings (correlation 0.96–0.99). The result is a property of the data, not an artifact of one technique.

Canonical sources

ONYX draws on the primary publishers of Nigerian economic data, supplemented by international references:

NBSCBNDMONUPRCNMDPRANCCNDICNPARMAFCIMFWorld BankOPEC/EIA

See it in the terminal

Intelligence an analyst can defend — answers and decisions, not data tables. Request a sample brief or enter the terminal.