Choosing the right WorldPop population data for you
We produce many different types of small area population datasets to meet a wide range of needs and applications. We also build tools and training materials to help you to construct your own datasets. This guide will help you find the right WorldPop data and tools to meet your needs..
On this page you will find information on:
Download our open population data
We construct two broad types of small area population dataset:
- If you are interested in recent small area population estimates for a specific country, country-specific datasets are likely to be the most accurate and relevant for your needs. These are often constructed in partnership with governments and UN agencies for a recent time-point and use input datasets and modelling methods that are typically bespoke for the focus country.
- If you are interested in small area estimates that can be compared over regional or global scales and multiple time periods, or temporal trends in a specific country, then datasets that cover all countries in the World over multiple years using a consistent modelling method are more likely to meet your needs.
Build your own small area population data
If you are familiar with modelling methods, our ‘Peanut Butter’ rapid mapping tool, PopRF R Code Package and training materials will enable you to build your own small area population count or density datasets.
If you want to familiarise yourself with our methods before using our tools and code packages, please review our datasets and tools overview, webinars, publications and GEO knowledge hub packages (top-down and bottom-up) for further information.
Small area data on population characteristics or dynamics
In addition to age and sex-structured small area population count and density data, we produce a variety of spatial datasets that provide small area estimates of the characteristics of residential populations, their demographics, and their mobility and migration patterns.
WorldPop small area estimates of population characteristics, development indicators and access to services are constructed from geolocated household survey data using Bayesian geostatistical modelling approaches that are similar to those used by the Demographic and Health Surveys program. These include the grid cell estimates of proportions of relevant populations that are vaccinated, literate or using healthcare. Examples of the production of poverty and births and pregnancies estimates are available, and you should also review our webinars and publications.
Small area metrics relating to population dynamics are also available. These include subnational migration flows, monthly population change estimates, air traffic flows, settlement growth and the locations and timings of holiday periods.
Choosing the right individual country population dataset
WorldPop produce many different types of small area population datasets for individual countries to meet a wide range of needs and applications. These often involve different trade-offs in modelling methods and input datasets that should be understood when you make choices on which dataset to use. Our gridded population estimates datasets and tools explainer offers an overview of our methods, and our webinars, publications and GEO knowledge hub packages (top-down and bottom-up) provide further details.
Population estimates for a recent time point
We provide a variety recent datasets using methodological approaches that are determined by the country and data input situation. You will find brief guidance on these below, but for a more in-depth background we recommend these papers:
- Small area population denominators for improved disease surveillance and response (in Epidemics)
- Spatially disaggregated population estimates in the absence of national population and housing census data (in PNAS)
- Advances in Small Area Population Estimation in the Absence of National Census Data (Preprint)
We develop and employ two broad modelling approaches in the production of small area population estimates for recent time periods. The ‘Top-down’ disaggregation and ‘Bottom-up’ estimation sections of our datasets and tools overview explain the differences between the two approaches and should be consulted to guide your dataset choice.
‘Bottom-up’ modelled population estimates
We co-develop bottom-up modelled estimates with governments and other partners. Countries where we have adopted this approach include Afghanistan, Burkina Faso, Cameroon, Colombia, the Democratic Republic of the Congo, Ghana, Guinea, Jamaica, Mali, Nigeria, Papua New Guinea, South Sudan and Zambia. These datasets, methodological details and associated uncertainty metrics can be downloaded from the WorldPop Open Population Repository (WOPR). Note that each dataset is produced using different bespoke modelling methods, tailored to the country context, data availability and national needs.
Because new data are regularly incorporated and improvements to models are made, different versions of each population dataset may be available. Also included within the WOPR are a collection of open datasets on buildings and boundaries, which were either used in the population model construction or derived using the output datasets.
‘Top-down’ disaggregated population estimates
We use top-down modelling methods across different projects, producing a range of different outputs that match to various needs and country contexts. Past projects include:
- Collaborations with the National Statistical Offices or Ministries of Health of Cameroon, Chad, Cote d’Ivoire, Guinea, Kenya, Mali, Mozambique, Niger, Senegal, Sierra Leone, Uganda and Zambia in constructing small area population datasets through disaggregation of either census counts or official projections to 100m grid squares. These are available through the WorldPop Open Population Repository (WOPR)
- Collaborations with the UNFPA Latin America and the Caribbean Regional Office (LACRO) and National Statistical Offices across the region to disaggregate official estimates and census counts to grid squares. Methods and data are available in High-resolution gridded population datasets for Latin America and the Caribbean using official statistics published in Scientific Data.
- Collaborations with UNFPA on the construction of small area population datasets (for Ukraine, Sudan, Türkiye, Yemen and others) through disaggregating Common Operational Datasets on Population Statistics (COD-PS).
- Top-down model datasets are available for all countries through our global data collection.
Because new data are regularly incorporated and improvements to models are made, different versions of each population dataset may be available. Also included within the WOPR are a collection of open datasets on buildings and boundaries, which were either used in the population model construction or derived using the output datasets.
Note that the bottom-up and top-down datasets described above are likely to be more reliable if recent estimates are required
Choosing the right global/regional or multi-temporal population dataset
Time-series of small area population data
Constructing multi-year time series of small area population estimates requires a range of methodological assumptions and trade-offs, producing datasets that are likely to be less accurate than those built for just a single, recent time point. We produce multi-year datasets and global mosaics for population counts, age/sex breakdown and population densities.
WorldPop’s 2015-2030 annual age and sex-structured population datasets are constructed through top-down disaggregation modelling of areal unit-linked population estimates derived from censuses and projections. They are available at 100m or 1km grid square scales, adjusted to match UN national population total estimates, and as per-grid cell totals, broken down by age/sex classes or as density estimates.
Additionally, the multi-temporal library of covariates used to construct the global datasets are accessible. The subnational age and sex-structure data assembled to construct these global layers can also be explored through the WorldPop Demographics portal.
Alternative recent estimates
In addition to the global data outlined above, a recent collaboration with the UNFPA Latin America and Caribbean Regional Office (LACRO) involved production of top-down modelled estimates for all countries in the LACRO region (published in Scientific Data), built from the most recent official government subnational estimates or census data. If your interest is in recent data for countries in the LACRO region that align with official government estimates, then these data may be more appropriate for use than those in our global collection.
