Meridian Intelligence · Methodology

How we validate every figure we publish.

Meridian Data turns licensed public statistics into source-traceable analyses. Public data is only useful when a reader can trust both the number and the interpretation around it. This page documents exactly how each analysis is sourced, checked, and published — and where our figures have limits.

The process, step by step

Source selection

Each analysis begins from one licensed public dataset. We do not aggregate or blend feeds into a single page. The dataset, its publisher, its licence, and the exact retrieval date are recorded in the post’s footer and structured metadata before any writing begins.

Figure extraction

Every headline number is pulled from a specific row of the source release, not paraphrased from secondary coverage. Each structured claim on a post links back to the source row (or row index) it came from, so any reader can verify the figure at its origin.

Claim scan-testing

Before publication, every quantitative claim is checked against the underlying value for direction, magnitude, and unit. Claims that cannot be tied to a source row are removed rather than published unverified.

Context and interpretation

The analysis adds the part public data does not: what the number means, how it compares to peers, what drives it, and where the figure can mislead. This interpretive layer — limitations included — is the original contribution of each piece.

Provenance stamp

Each published analysis carries a run ID, publication date, source licence, and attribution. The pipeline status of the publication system is public, so readers can see when and how each piece was produced.

Where our data comes from

We work only with open or licensed statistical sources and cite each one on the analysis that uses it. The most frequent sources:

Source	How we use it	Licence
World Bank Open Data	Macroeconomic, demographic, health, and energy indicators by country and year.	CC BY 4.0
OECD Statistics	Cross-country comparative benchmarks for the peer-set panels in premium briefs.	OECD Terms
Our World in Data	Long-run series and standardised emissions / energy datasets.	CC BY 4.0
IMF & MoSPI	Inflation, national accounts, and India-specific official statistics.	Source terms

Limitations & corrections

Public data has known limits. Official statistics are revised, lag the present by months or years, and differ in definition across countries. Where a figure is sensitive to methodology — life expectancy to infant mortality, labour-force rates to how participation is counted — we say so in the analysis rather than presenting the number as settled.

Cross-country comparisons are made carefully. When a peer comparison would mislead — because reporting standards differ or peers don’t report an indicator — we either note the gap or omit the comparison.

We correct in place. If a published figure is later found to be wrong or a source issues a revision, we update the analysis and note the change. To report an error, email hello@meridianintelai.com with the post URL and the figure in question.