Connecting population with place
Knowing how many people live in a country is an important starting point. Planning schools, assessing healthcare access or studying flood exposure also requires information about how people are distributed within that country.
WorldPop produces population datasets that represent these distributions through geographic grids. Its top-down modelling approaches combine population counts for administrative areas with geospatial information to estimate how those populations are distributed across smaller cells. These estimates provide a way to connect population information with other geographic data. WorldPop modelling overview
This article focuses on the 2020 constrained, UN-adjusted population datasets used in a project involving Bhutan. They belong to WorldPop’s archived Global1 collection and include total population estimates and separate age–sex layers. The grids have a resolution of 3 arc-seconds, approximately 100 metres at the equator. Global1 population data
What do the files contain?
The data are supplied as GeoTIFF rasters: geographic images in which each cell stores a numerical value. In a population-count layer, that value represents the estimated number of people within the cell.
Age–sex layers describe particular population groups. The Global1 collection separates male and female populations into age bands, mostly covering five years, with additional detail for the youngest ages. This allows researchers to examine groups relevant to a particular question, such as children, women aged 15–49 or older adults. Age–sex data overview
Two labels help explain the selected version. Constrained means that population allocation is restricted to cells identified as containing settlements. UN-adjusted means that national totals are adjusted to the corresponding United Nations population estimates; this 2020 collection uses the 2019 revision of World Population Prospects. Constrained modelling, 2020 UN-adjusted age–sex collection
From population grids to a research table
A practical workflow developed for the Bhutan data brings the total population raster and age–sex layers into a single table. It involves:
Reading valid cells from the total population raster.
Recording each cell’s centre coordinates and population estimate.
Adding values from the corresponding age–sex layers.
Exporting the combined information to CSV for further analysis.
The resulting table can contain country, year, coordinates, total population and separate age–sex columns. Researchers can then combine compatible age bands into broader groups. For example, the female bands from 15–19 through 45–49 can be summed to estimate women aged 15–49.
This workflow makes the data easier to use in statistical software and GIS. The coordinates identify grid-cell centres, rather than individual people or household addresses.
What could researchers investigate?
Population grids can support several kinds of analysis when combined with suitable external data:
Education: compare the distribution of children in relevant age groups with school locations or catchment areas.
Healthcare: estimate populations within service areas and examine access for particular demographic groups.
Disaster exposure: overlay population estimates with flood or other hazard maps.
Planning: compare settlement patterns and potential demand for infrastructure or public services.
These are possible applications. Travel times, service capacity and exposure require additional information beyond the population grids.
Using the estimates carefully
WorldPop values are modelled estimates. Fine grid resolution does not guarantee equally precise information about every cell, and results depend on the population inputs and the accuracy of settlement mapping. Missing settlements can affect where people are allocated. Modelling assumptions and limitations
When combining layers, researchers should check the year, version, coordinate system, grid alignment and age definitions. Missing-data values should be handled separately from zero population. Population counts should also retain their decimal precision during processing; rounding every cell before aggregation can distort totals.
Where to begin
The Global1 population archive and age–sex archive provide starting points for locating the products discussed here. Each dataset’s metadata and recommended citation should accompany its use.
WorldPop offers a practical connection between population statistics and geographic questions. Understanding its structure and assumptions helps researchers turn population grids into evidence about services, exposure and differences between places.