WorldPop is an open, gridded population and demographic data repository produced by a research group based at the University of Southampton’s School of Geography and Environmental Sciences, initiated in October 2013 to combine census records, household surveys, satellite imagery, cellphone data, and other geospatial sources into high-resolution estimates of population distribution and characteristics. Its datasets are used across public health modeling, epidemiology, disaster response and humanitarian planning, and broader demographic research wherever a study needs to know not just how many people live somewhere, but where, at a finer spatial resolution than census units alone provide.
This guide covers what WorldPop actually is, what kinds of research questions its data is suited to, its data formats and access model, its licensing terms, and how to cite a WorldPop dataset correctly in a publication or data management plan. It is scoped to the research-data-access angle specifically — researchers evaluating or using WorldPop data — rather than a general organizational profile.
What WorldPop is, and what problem it solves
Traditional population data is reported at the resolution of administrative units — countries, provinces, districts, census tracts — which is often too coarse for research questions that depend on where people actually are within those boundaries. WorldPop addresses this by producing gridded population data: raster datasets that estimate population counts or densities within regular grid cells (commonly around 100-meter and 1-kilometer resolution, depending on the dataset and country), rather than as a single figure for an entire administrative area.
The underlying method, sometimes described as “top-down” or “bottom-up” small-area population estimation, disaggregates census totals across a landscape using ancillary geospatial covariates — settlement extent from satellite imagery, land cover, road networks, nighttime lights, and similar predictors — to model where within a census unit the population is actually concentrated. WorldPop’s methodology and the resulting open dataset were formally described in a peer-reviewed paper, Tatem, A.J. “WorldPop, open data for spatial demography,” Scientific Data 4, 170004 (2017), which remains the standard methodological reference for the project.
What kinds of research WorldPop data supports
Because WorldPop expresses population as a continuous spatial surface rather than a table of administrative totals, it is used as an input layer across several distinct research contexts rather than as an end product in itself:
- Public health modeling and epidemiology. Disease burden estimates, intervention coverage targets (e.g., vaccination campaign planning, bed-net distribution), and outbreak or transmission models frequently need denominator populations at a finer spatial scale than administrative reporting provides — for example, estimating how many people live within a given travel time of a health facility, or the population at risk within a modeled flood or vector-habitat zone.
- Disaster response and humanitarian planning. Agencies and researchers estimating populations exposed to or displaced by a specific hazard (earthquake, flood, cyclone) use gridded population layers to produce rapid exposure estimates before updated on-the-ground figures are available, and to plan the scale of a response.
- Demographic and development research. Studies of urbanization, migration, accessibility (e.g., travel time to services), and small-area poverty or development indicators often pair WorldPop’s population grids with other spatial covariates as a standard input layer.
- Age- and sex-structured analysis. Beyond total population counts, WorldPop also produces age- and sex-disaggregated gridded datasets for many countries, which matter for research questions where risk or need varies by demographic group (e.g., maternal health, school-age population estimates).
In each case, WorldPop functions as a data source or input layer for a downstream analysis, not as a repository of finished research findings — this is a useful distinction to keep in mind when describing a data source in a data management plan (DMP) or methods section.
Data formats and access model
WorldPop datasets are distributed primarily as gridded raster files, most commonly GeoTIFF, which is directly compatible with standard GIS software (QGIS, ArcGIS, R’s raster/terra packages, Python’s rasterio/GDAL stack) without conversion. Datasets are organized by country and by theme — population counts, age/sex structure, births, and other derived layers — and are browsable through WorldPop’s online Data Catalogue and the WorldPop Hub, with individual datasets also findable and downloadable via the WorldPop Open Population Repository (WOPR) and, for some products, through partner platforms.
Access itself does not require registration, an account, or a data use agreement for the openly licensed products — a researcher can locate a dataset in the catalogue and download it directly. This open, no-gatekeeping access model is a meaningful factor in WorldPop’s suitability for time-sensitive uses such as disaster response, where a data sharing agreement negotiation would be impractical.
Licensing
WorldPop datasets are released under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Under CC BY 4.0, reuse, adaptation, and redistribution — including commercial use — are permitted provided the source is properly attributed; it does not require that derivative works be shared under the same license (unlike a CC BY-SA “share-alike” license) or restrict commercial use (unlike CC BY-NC). For background on how CC BY compares with more restrictive Creative Commons variants, see CASRAI’s CC BY vs. CC BY-SA comparison and the Creative Commons License dictionary entry.
Because the license is applied uniformly and does not require negotiating individual terms per dataset, WorldPop functions in practice like other openly licensed domain repositories in this respect — a researcher can plan a data management plan around it without a separate data access agreement step, unlike sensitive-data repositories that gate access behind review.
How to cite WorldPop data
WorldPop’s own citation guidance has two parts, and both should generally be followed together:
- Cite the specific dataset. Each dataset in the WorldPop catalogue is registered with its own citation and, for many datasets, its own DOI — the catalogue page for the dataset you downloaded provides the exact recommended citation string and DOI to use. Cite the specific dataset version and DOI you actually used, not a generic reference to the WorldPop project, since datasets are periodically updated and reissued (e.g., annual population estimate updates for a given country).
- Cite the WorldPop project itself as the source. WorldPop asks that, alongside the dataset-specific citation, the WorldPop website (www.worldpop.org) be cited as the general source. Many published uses of WorldPop data also cite the founding methodology paper (Tatem, 2017, Scientific Data) where a description of the underlying modeling approach is relevant to the citing work.
This two-part pattern — cite the specific dataset/DOI, plus the parent project — is consistent with how CASRAI generally recommends citing data drawn from any repository; see the Repository Citation dictionary entry and CASRAI’s Data Publication Practices guide for how dataset citation fits alongside data papers and repository deposits more broadly.
Where WorldPop fits relative to other population and repository sources
WorldPop is a domain-specific data producer and repository focused specifically on gridded population and demographic data, distinct from generalist repositories such as Zenodo or Dryad (where researchers deposit heterogeneous datasets from any field) and distinct from a national statistical office’s own census releases (which WorldPop uses as an input, but redistributes at a different spatial resolution and modeling basis). For researchers evaluating repository options more broadly as part of a DMP, see CASRAI’s How to Choose an Open Data Repository guide and the Domain repository and Discipline-specific repository dictionary entries for how a domain repository like WorldPop differs from generalist alternatives.
Because it produces modeled estimates rather than raw survey microdata, researchers using WorldPop outputs should document the specific dataset version, resolution, and modeling method used (available on each dataset’s catalogue page) in their methods section — this is standard practice for any derived/modeled geospatial data layer and supports reproducibility in line with FAIR Data Principles.
Frequently asked questions
Is WorldPop data free to use?
Yes. WorldPop’s datasets are openly licensed under CC BY 4.0 and can be downloaded from the WorldPop Data Catalogue and WorldPop Hub without a fee, registration, or data use agreement.
Can WorldPop data be used commercially?
Yes, subject to attribution. CC BY 4.0 permits commercial use and adaptation; it does not require share-alike redistribution of derivative works.
What resolution is WorldPop’s population data?
Resolution varies by dataset and country, but WorldPop’s gridded population products are commonly produced at approximately 100-meter and 1-kilometer grid-cell resolutions, considerably finer than typical administrative-unit reporting.
Does WorldPop provide age- and sex-disaggregated data, or just total population counts?
Both. In addition to total population count and density grids, WorldPop produces age- and sex-structured gridded datasets for many countries, which support research questions where a demographic subgroup’s population matters (e.g., women of reproductive age, school-age children).
How do I cite a WorldPop dataset in a paper?
Cite the specific dataset (using the citation and DOI provided on that dataset’s catalogue page), cite www.worldpop.org as the general project source, and, where the underlying methodology is relevant, cite Tatem, A.J. (2017), “WorldPop, open data for spatial demography,” Scientific Data 4, 170004.
Is WorldPop the same as census data?
No. WorldPop uses census totals and other data sources as inputs to a spatial disaggregation model, producing gridded estimates of where population is likely distributed within administrative units — it is a modeled, derived product, not a replacement for a country’s official census release.
