WorldPop dataset
1. Introduction to WorldPop dataset
WorldPop dataset official website: https://www.worldpop.org/
Launched in 2013, WorldPop unites the continent-focused AfriPop, AsiaPop and AmeriPop projects to produce detailed and freely available maps of population distribution and composition for the entirety of Central and South America, Africa and Asia. WorldPop estimates the number of people, population density, etc. per 100m x 100m
grid square on Earth . Population spatial databases have been widely used in disease burden estimation, epidemic modeling, resource allocation, disaster management, accessibility modeling, transportation and urban planning, poverty mapping, and environmental impact assessment, etc. WorldPop is a 100-m resolution grid population estimate using a custom approach ("top-down" or "bottom-up") , and uses both constrained and unconstrained decomposition methods. The WorldPop top-down modeling approach takes a global database of administrative-unit-based censuses and projected counts for each year 2000-2020 and decomposes them into grid-unit-based counts using a detailed set of geospatial datasets. Two approaches have been taken, generated in several countries using the random forest machine learning approach described in Stevens et al., and the code is provided.
- Estimates of squared grids of all land in the world ( unconstrained );
- Estimates are made only within areas mapped to contain established settlements ( constrained ).
A recent comparison of constrained versus unconstrained top-down mapping for multiple countries in Reed et al. and Stevens et al. found limited differences in the accuracy of national maps produced by each country, but significant differences between countries. difference.
The difference between Top-down unconstrained and Top-down constrained:
ie: Unconstrained means that the method produces a non-zero distribution of population to all land grid cells, leading to misallocation of population to uninhabited areas and underestimating the urban population in some areas. Whereas constraints mean that in cases where settlements/buildings are accurately mapped, the output shows a more accurate population distribution, with low predicted populations in potentially uninhabited areas.
It is recommended to use Top-down constrained data!
Top-down constrained data on the spatial distribution of China's population in 2020: https://hub.worldpop.org/geodata/summary?id=49730
2. WorldPop dataset download
Data address: https://hub.worldpop.org/
This page has various types of data such as population size and population density. You can download the required data directly, but the speed is relatively slow.