Source
All figures are from the United Nations Department of Economic and Social Affairs —
World Population Prospects 2024,
the standard reference used by national statistical offices, the World Bank, IMF, OECD,
and academic researchers. Historical figures (1950–2023) come from country censuses and vital
registration; values from 2024 onward use the UN’s
medium variant projection.
Pre-1950 segments use the HYDE 3.3 long-run reconstruction (relevant only if extended).
Data is loaded from
demographics.json — built once from the public
OWID dataset
that merges WPP and HYDE into a single time series.
Why we do not roll our own projection
Demographic momentum is driven by current age structure, not by past growth rates.
A naive extrapolation (linear or compound growth) will miss every interesting case —
China’s decline, Japan’s shrinkage, Sub-Saharan Africa’s acceleration. Cohort-component
models that track births, deaths, and migration by age are the only credible approach,
and the UN’s team has been refining them since 1951. We use their result.
What “share of world” means
For every year shown, we sum the population of all countries with data and divide each
country’s population by that total. This is not the official UN world total
(which includes a few aggregates we exclude), but it differs by less than 0.1%
and lets the colors stay internally consistent.
Color scale
The map uses a six-step bucketing on percent-of-world —
0.1%,
0.5%,
1%,
5%,
10%,
20% — chosen so the long tail
of small countries is visible while still distinguishing China and India from the next pack.
Bubble view
Each circle’s
area (not diameter) is proportional to population, and bubbles are
laid out by a deterministic spiral pack so the same country sits in the same neighborhood
across years — making movement legible as you drag the slider.
Limitations
• Medium-variant projections carry uncertainty bands that widen the farther out you look.
The UN also publishes a low and a high variant; we show only the medium for clarity.
• Country borders shown are the present-day topology; we apply current borders to historical
populations. South Sudan, for example, appears as a country in 1950 even though it did not
exist politically — the population data belongs to the area, not the state.
• A few small territories with no WPP entry (e.g. Vatican City) appear as “no data.”