Environmental · geography & country data

World Atlas

Every country, two projections, and every number sortable.
countries with data
drag to pan · scroll to zoom · click a country
Equal Earth — every country at its true relative size.
Shades the map, sorts the table and drives the top twenty below. Countries are shaded by rank, not raw value — one outlier would flatten the rest.
Everything else fades back.
Selected

Pick a country

click the map, or search above
A reference atlas, not a political statement. Boundaries follow Natural Earth and figures follow the World Bank; both have editorial positions on disputed territory that this page does not adjudicate. Eight landmasses on the map carry no data because no World Bank country record exists for them — they are drawn grey and named.
Browse

Every country, every metric, sortable

click a name to put it on the map
Ranking

Top 20

Same metric as the map, both ends of it. The continent chips above filter this too.
Groups

What each bloc adds up to

click a card to highlight it on the map
Shares are of the world totals in this dataset. A country in several groups is counted in each.
Categories

Three relationships worth a chart

Every dot is a country, coloured by continent; the one you have selected is ringed. Hover a dot for its figures, click it to select it, or click a continent in the key to pull it out of the crowd. These are the three shapes a table of this data hides.

Money and yearsGDP per person (log) against life expectancy
This is the Preston curve, and on a log income axis it is strikingly close to a straight line: the fitted slope drawn here is , or roughly ten years of life expectancy per tenfold increase in income. Which is the diminishing return, stated exactly — going from $1,000 a head to $10,000 buys ten years for $9,000, and the next ten years costs $90,000. The curve is not bending; the money is. Note too how far countries sit above and below the trend at the same income: GDP per person is not a proxy for how well a country lives.
Room to liveland area against population, both log
On log–log axes, equal density is a straight diagonal, and the dashed one is the world average of . Distance above the line is crowding and below it is space, so the whole spread of human density — five orders of magnitude — is one glance.
Clock and sunhow far each capital’s clock runs from its own sun
Solar noon is geometry, so the interesting number is not the offset but the gap: legal time minus the sun over the capital. Zero is the dashed line and the shaded band is an hour either side. Sixteen countries sit outside it and every one is ahead of its sun, never behind — which is what you would expect if clocks are set for evening daylight rather than for noon. Plotting the raw offset instead makes a neat diagonal staircase and hides all of this.
Time

Clocks are a decision, not a measurement

Solar time is geometry: the sun is overhead at longitude ÷ 15 hours from UTC, exactly. Legal time is a choice, and the two come apart by more than an hour in sixteen countries.

Legal offsets are standard time from the IANA time-zone database, taken as the lower of the January and July offsets so that summer time is excluded in both hemispheres. Solar offset is the capital’s longitude. Convert an actual time between places →
Play

Top Trumps

pick a statistic

Choose both countries, then click the statistic you think the left-hand card wins on. Deal a random pair instead and the right-hand card stays face down until you commit to a stat.

Only countries with all six figures can be dealt.
Face down until you pick a stat.
“Use selected” takes the country you picked on the map.
Field notes

Reading a world map without being fooled by it

How it works

Two projections, and why they disagree

You cannot flatten a sphere without breaking something

Every world map is wrong in a way its maker chose. A projection can preserve area, or angles, or distance from one point — never all three, and area and angle never together. That is not a limitation of cartography; it is a theorem. Gauss proved in 1827 that a sphere has intrinsic curvature and a plane does not, so no mapping between them can preserve both shape and size.

So the honest question is not which map is right. It is which distortion you can live with for the job in front of you.

Equal area: Equal Earth

The default here is Equal Earth (Šavrič, Patterson and Jenny, 2018). Every country occupies its true share of the page, so comparing sizes by eye actually works. It buys that by shearing shapes, most visibly at the poles.

This is the modern answer to the argument the Gall–Peters projection made loudly in the 1970s — that Mercator shrinks the tropics. Equal Earth makes the same point without Gall–Peters' badly stretched continents.

One correction worth making: the United Nations emblem is often cited as the "true size" map. It is not equal-area. It is an azimuthal equidistant projection centred on the North Pole, which preserves distance from the centre and distorts area severely toward the edges — Antarctica becomes the entire rim.

Conformal: Mercator

Mercator (1569) preserves angles, which is why it has never been replaced for navigation: a constant compass bearing is a straight line on it. The price is area, and the price is steep — the scale factor is 1/cos(latitude), so at 60° everything is drawn twice its true size and at 75° nearly four times.

Switch between the two and watch Greenland. It is genuinely about the size of Algeria, and Mercator draws it larger than Africa. Mercator is also the taller map: it is clipped here at 80° because it runs to infinity at the poles, and even clipped it needs half again the height of Equal Earth to fit the same world.

The date line is bookkeeping, and it will streak your map

Natural Earth stores Russia, Fiji and Antarctica as rings that run right up to the antimeridian and continue on the far side at the opposite longitude. Draw those points in order and the pen jumps from +180° to −180° — straight across the whole map. Four rings in this dataset do it, and the streaks they leave look exactly like a rendering bug because that is what they are.

So every ring is cut wherever consecutive points jump more than 180° of longitude, and each piece is closed on its own. A piece whose two ends land on opposite edges — Antarctica, which genuinely wraps the globe — is closed through the pole instead, which is why the ice fills the bottom of the map rather than being sliced off flat.

Where the numbers come from

Population, land and surface area, GDP, GDP per capita, life expectancy, urban share, CO₂ per head and forest cover come from the World Bank Open Data API, each taken at its most recent year available for that country — so the year is carried alongside the figure rather than assumed. Highest and lowest points, ISO codes and continents come from Wikidata. Boundaries are Natural Earth at 1:110m.

Nothing here is estimated to fill a gap. A country with no reported figure shows a dash, because a blank is honest and an interpolation is not.

Two things the land-area column is not. It is land, so it excludes inland water — and for Greenland it excludes the ice sheet, which is why it reads 410,450 km² rather than the 2.17 million you may have in mind. The World Bank also publishes a "surface area" series; it is not shown here, because it carries marine area and puts Bermuda at 4,290 km² against a real 54. A column that wrong does not belong in a sortable ranking.

What the map cannot show

At 1:110m resolution the world is 177 shapes, and the smallest states are not among them — Singapore, Malta, Monaco, Bahrain and two dozen others are real countries with real data that are simply too small to draw at this scale. They appear in the browser and the rankings and not on the map, and the counter above the map says how many.

Eight drawn landmasses have no World Bank country record: Western Sahara, the Falklands, the French Southern Territories, Taiwan, Antarctica, Northern Cyprus, Somaliland and Kosovo. Each is contested, uninhabited, or both. They are drawn, greyed and named rather than deleted.

Common questions

Questions this raises

Which projection is the correct one?

Neither, and the question has no answer. Equal Earth is correct for comparing sizes; Mercator is correct for holding a bearing. A map that claimed to be correct for both would be making a mathematically impossible claim.

Why does the population not match the figure I know?

Because it is dated. Each figure carries the year the World Bank last reported it for that country, and those years differ — a 2019 figure for one country sits next to a 2024 figure for another. The year is shown next to the number for exactly that reason.

Is GDP comparable between countries?

Only roughly. These are current US dollars at market exchange rates, so a currency movement changes a country's GDP without anything happening in its economy. Purchasing power parity tells a different and often more useful story, and is not shown here.

Version history · 2 releases
  1. v5.122026-09-12Atlas gains time-zone linking and three category charts — and the model caught the Preston curve prose describing the wrong axis
  2. v5.132026-09-12The atlas rebuilt: the streak across Siberia was the date line, the map is full width with the controls under it, and the charts got a key, a hover and a point

Releases in which this page changed, newest last. Derived from the archived copy of every release, not from notes written afterwards — so it reflects what actually shipped. Site-wide passes are left out; they are in the full changelog.