How the figures here are put together

The UN publishes its population on two dates, and this site uses both. A mid-year figure is for 1 July: it is the one to use when you want to know how many people were alive during a year, and it is what every pyramid on the site is drawn from. A 1 January figure is the one the UN attaches its ranges of uncertainty to, so the charts that show a range use that date instead.

Estimates run to 2023. Every year after that is the UN's medium projection, which is the middle of the range it thinks likely rather than a forecast anyone guarantees. Projected years are drawn with a hatch everywhere on this site, and a projected figure is never given without its range where the UN publishes one. Countries have ranges; regions and groups, in this release, do not, and a page that cannot draw one says so.

One more thing you will meet on a place page: the population added up age by age can differ a little from the published total. The UN rounds each age separately, and for a small place those roundings add up to something you can see. Where they do, the page prints both numbers rather than picking one.

A city here is not a city government's area. It is the UN's Degree of Urbanization city: a cluster of square kilometres with at least 1,500 people living on each one, and at least 50,000 people in the cluster altogether. The UN draws it from satellite pictures of buildings (the Global Human Settlement Layer) and from census maps, the same way in every country. Its edge follows the built-up land, so the number can be much larger, or much smaller, than the one the city publishes about itself.

Because the rule is new, these figures are not comparable with the UN's older city lists (World Urbanization Prospects 2018 and before). Nor are they comparable with a national definition of "urban", which is a different thing in every country.

The UN estimates every city once every five years, 1975 to 2050. The years in between are its straight line between two estimates, so these pages show the five-yearly figures and nothing between them. After 2025 they are projection, with no range of uncertainty. A place joins the list the first time it passes 50,000 people, and leaves it when it falls below that or grows into a neighbouring city. The UN does not say which of the two happened.

"How sure the UN is" is the UN's own plausibility rating for each figure. It comes down to how recent and how detailed the census maps behind that figure are, which is why cities in a country that has not run a census lately are rated low.

A city's share of its country is our own division: the city's people divided by the UN's count of everyone in that country the same year. Added up, the cities agree with the UN's national city totals to within 0.6% in every year and exactly in 2025; in some earlier years a country's list is missing a small city that the national total already counts. Where the UN gives a city's name only in its own script, a romanization is used here instead, which applies to 8 of them. The three cities that reach across a border (Macao, Gibraltar and Monaco) are never drawn as a share of one country, because each holds more people than one side of the border does.

Only cities that reach 300,000 people in some year have a page of their own. That is why a ranking here skips numbers: a gap is a smaller city.

A comparison puts two places on one set of axes and one slider, so every picture on the page is at the same moment for both. 250 pairs have a page of their own; any other two places can be compared from the boxes at the top of one.

The age half the people are older than, which is the median age, is worked out from the UN's population by single year of age, exactly as it is on each place's own page. "If each place were 100 people" rounds the three age groups to whole people so that they add up to 100, which is why a group can read one person high or low against its exact share.

The pictures that show every country at once count the 236 countries and areas the UN reports. Regions, continents and groups are never counted in them: they are made of those same countries, and counting both would count many people twice.

Two measures on those pages are easy to misread. Children per woman, the total fertility rate, is how many children a woman would have if she lived through one year's birth rates at every age of her life. It is not the number of children the women alive today have had. Life expectancy at birth is how long a newborn would live if that year's death rates never changed, and it is an average: plenty of people live much longer than it, and it is pulled down by anyone who dies young.

Verified 2026-09-30 · 73 canaries