Method

Evidence before recommendation.

The site is generated from reviewed observations. Presentation never changes the geographic level, source date or confidence of the underlying evidence.

Scope stays visible

Country, region, metro and municipality figures are labelled as such.

Absence is a real state

Not collected, unavailable, stale and collection failure are not replaced by guesses.

Citations point to evidence

References identify the dataset edition, table, query or document behind a claim.

Judgement is labelled

Computed comparison and reviewed editorial judgement are kept visibly distinct.