Methodology

1. What the index is

The Country Standard Index is a single 0–100 score per country per year that summarises quality of life along 7 components: affordability and income, health and healthcare access, jobs, education, safety, freedom, and environment and infrastructure. It is built only from published statistics (28 indicators), with no survey of opinion run by us and no hand-adjusted values.

It is designed to be decomposed. Every country page shows the index, the contribution of each component in index points, the indicators inside each component, and the raw official series with the year each value was observed.

2. Structure and weights

Index = Σ (w_c × component_c) ÷ Σ w_c over the components that have data (at least 5 of 7 are required). Component = simple mean of its available indicator scores. Default weights are equal; there is no official basis for weighting one dimension of life above another, so the ranking page lets you set your own.

3. Scoring each indicator

1. Transform. Income, infant mortality, homicide and PM2.5 are log-transformed so that differences at the low end count more than the same absolute difference at the high end. Inflation below zero is set to zero; gross tertiary enrolment is capped at 100%.
2. Goalposts. For each indicator the worst and best goalposts are the 2.5th and 97.5th percentiles of all observed country-years from 2000 to 2024, pooled. They are the same for every year, which is what makes history meaningful: a country's score moves only when its own numbers move. The 2.5/97.5 trim stops one extreme country-year (a hyperinflation, a war) from compressing everyone else into a narrow band.
3. Score. 100 × (value − worst) ÷ (best − worst), clipped to 0–100, with the direction flipped for indicators where lower is better.

Every indicator page spells these steps out for that series (native scale, transform, goalposts in both raw and transformed units, flip, clipping), and every value on a country page or indicator table can be opened to show the arithmetic applied to that exact number. The build is checked by npm run data:verify, which recomputes all ~124,000 stored scores from the raw values and the published goalposts and confirms each indicator is monotone in the stated direction.

4. Missing data

• Carry-forward. If a year has no published value, the most recent observation is carried forward for at most maxCarry years (2–10 depending on how often the source publishes; see table). Every carried value is flagged on the country and indicator pages, and each score shows what share of its inputs is carried.
• Missing indicators. A component is computed from whichever of its indicators are available; the country page shows “n/m indicators”. A country needs at least 5 components to get an index, and at least 70% of all indicators to be ranked. Countries below that threshold still have a score, shown with a warning, and can be included from the ranking page.
• Nothing is imputed from other countries. We never fill a country's gap with a regional average; the gap stays visible.

5. Indicators and sources

Every series links to two places: downloaded from is the exact URL our pipeline fetched (the closest point to the number on this site), and built from is the chain of organisations behind it. Where a value is assembled from several sources, each entry says what that source contributes; the indicator page also shows the data provider's own source statement verbatim, and every country row has a link to the series filtered to that country.

Type: direct measures the thing itself; proxy stands in for something not measured consistently across countries. Source type: statistical = counts, registers or household surveys compiled by statistical agencies; modelled = statistical estimates (ILO, UN IGME, GBD) that fill gaps between reported years; expert = assessments aggregated from expert and survey sources (WGI, V-Dem).

Not scored but shown for cross-checking: Population, total (World Bank WDI (UN World Population Prospects, national statistical offices, Eurostat)); Human Development Index (UNDP) (UNDP Human Development Report 2025).

6. Where official numbers are missing, and proposed proxies

7. Known limitations

• Equal weights are a choice, not a finding. Use the sliders to test how sensitive the ranking is.
• Governance and freedom indicators are expert assessments, not counts. They are the standard sources (used by the UN, EU and World Bank) but carry a standard error; the WGI publish it.
• Small states have thin data and are easily unranked; this is a data limitation, not a judgement about them.
• Carry-forward keeps recent years comparable but means that a country's latest score can be partly last year's. The carried share is always shown.
• National averages hide inequality within countries; the Gini index is the only distributional measure included.

8. Reproducing the data

npm run data:fetch   # pull all series from the World Bank API, UNDP and OWID into data/raw/
npm run data:build   # compute scores, components, index, ranks → data/*.json
npm run build        # build the site

The indicator catalogue (ids, directions, transforms, carry limits, sources) is a single file, scripts/indicators.mjs; add a row there and re-run to extend the index.