The world's most accurate population datasets. Seven maps/datasets for the distribution of various populations in Zimbabwe: (1) Overall population density (2) Women (3) Men (4) Children (ages 0-5) (5) Youth (ages 15-24) (6) Elderly (ages 60+) (7) Women of reproductive age (ages 15-49). Methodology These high-resolution maps are created using machine learning techniques to identify buildings from commercially available satellite images. This is then overlayed with general population estimates based on publicly available census data and other population statistics at Columbia University. The resulting maps are the most detailed and actionable tools available for aid and research organizations. For more information about the methodology used to create our high resolution population density maps and the demographic distributions, click here. For information about how to use HDX to access these datasets, please visit: https://dataforgood.fb.com/docs/high-resolution-population-density-maps-demographic-estimates-documentation/ Adjustments to match the census population with the UN estimates are applied at the national level. The UN estimate for a given country (or state/territory) is divided by the total census estimate of population for the given country. The resulting adjustment factor is multiplied by each administrative unit census value for the target year. This preserves the relative population totals across administrative units while matching the UN total. More information can be found here

The population of the world, allocated to 1 arcsecond blocks. This refines CIESIN’s Gridded Population of the World project, using machine learning models on high-resolution worldwide Digital Globe satellite imagery. More information.

There is also a tiled version of this dataset that may be easier to use if you are interested in many countries.

This dataset comprises a series of five land use and land cover (LULC) maps of western Murewha District, Zimbabwe, spanning the years 2002, 2007, 2013, 2018, and 2023. The overall accuracy scores for these maps are 0.93, 0.91, 0.90, 0.90, and 0.90, respectively. These maps were generated using open-access Landsat satellite imagery (30m resolution) from Landsat 5, 7, and 8, enabling consistent spatial resolution and temporal coverage. Each map integrates two images from the crop/wet and dry seasons, ensuring comprehensive seasonal representation. Key radiometric indices (NDVI, RVI, NDWI2, BI2) and a 30m resolution DEM were applied for enhanced classification accuracy. The algorythm used for the classification is a pixel random forest using Python 3.7.4 and the library sklearn. The study focuses on wards within Chitopi and Mushaninga sub-districts.

Population density per pixel at 100 metre resolution. WorldPop provides estimates of numbers of people residing in each 100x100m grid cell for every low and middle income country. Through ingegrating cencus, survey, satellite and GIS datasets in a flexible machine-learning framework, high resolution maps of population counts and densities for 2000-2020 are produced, along with accompanying metadata. DATASET: Alpha version 2010 and 2015 estimates of numbers of people per grid square, with national totals adjusted to match UN population division estimates and remaining unadjusted. REGION: Africa SPATIAL RESOLUTION: 0.000833333 decimal degrees (approx 100m at the equator) PROJECTION: Geographic, WGS84 UNITS: Estimated persons per grid square MAPPING APPROACH: Land cover based, as described in: Linard, C., Gilbert, M., Snow, R.W., Noor, A.M. and Tatem, A.J., 2012, Population distribution, settlement patterns and accessibility across Africa in 2010, PLoS ONE, 7(2): e31743. FORMAT: Geotiff (zipped using 7-zip (open access tool): www.7-zip.org) FILENAMES: Example - AGO10adjv4.tif = Angola (AGO) population count map for 2010 (10) adjusted to match UN national estimates (adj), version 4 (v4). Population maps are updated to new versions when improved census or other input data become available.

The dataset is a zip file that contains 28 cloud optimized tiff files that cover the continent of Africa. Each of the 28 files represents a region or area - these are not divided by country. These 28 tiff files represent 2015 population estimates. However, please note that many of the country-level files include 2020 population estimates including: Angola, Benin, Botswana, Burundi, Cameroon, Cabo Verde, Cote d'Ivoire, Djibouti, Eritrea, Eswatini, The Gambia, Ghana, Lesotho, Liberia, Mozambique, Namibia, Sao Tome & Principe, Sierra Leone, South Africa, Togo, Zambia, and Zimbabwe. To create the high-resolution maps, machine learning techniques are used to identify buildings from commercially available satellite images then general population estimates are overlaid based on publicly available census data and other population statistics. The resulting maps are the most detailed and actionable tools available for aid and research organizations.

Social distancing is a public health measure intended to reduce infectious disease transmission, by maintaining physical distance between individuals or households. In the context of the COVID-19 pandemic, populations in many countries around the world have been advised to maintain social distance (also referred to as physical distance), with distances of 6 feet or 2 metres commonly advised. Feasibility of social distancing is dependent on the availability of space and the number of people, which varies geographically. In locations where social distancing is difficult, a focus on alternative measures to reduce disease transmission may be needed. To help identify locations where social distancing is difficult, we have developed an ease of social distancing index. By index, we mean a composite measure, intended to highlight variations in ease of social distancing in urban settings, calculated based on the space available around buildings and estimated population density. Index values were calculated for small spatial units (vector polygons), typically bounded by roads, rivers or other features. This dataset provides index values for small spatial units within urban areas in Zimbabwe. Measures of population density were calculated from high-resolution gridded population datasets from WorldPop, and the space available around buildings was calculated using building footprint polygons derived from satellite imagery (Ecopia.AI and Maxar Technologies. 2020). These data were produced by the WorldPop Research Group at the University of Southampton. This work was part of the GRID3 project with funding from the Bill and Melinda Gates Foundation and the United Kingdom’s Department for International Development. Project partners included the United Nations Population Fund (UNFPA), Center for International Earth Science Information Network (CIESIN) in the Earth Institute at Columbia University, and the Flowminder Foundation.

Social distancing is a public health measure intended to reduce infectious disease transmission, by maintaining physical distance between individuals or households. In the context of the COVID-19 pandemic, populations in many countries around the world have been advised to maintain social distance (also referred to as physical distance), with distances of 6 feet or 2 metres commonly advised. Feasibility of social distancing is dependent on the availability of space and the number of people, which varies geographically. In locations where social distancing is difficult, a focus on alternative measures to reduce disease transmission may be needed. To help identify locations where social distancing is difficult, we have developed an ease of social distancing index. By index, we mean a composite measure, intended to highlight variations in ease of social distancing in urban settings, calculated based on the space available around buildings and estimated population density. Index values were calculated for small spatial units (vector polygons), typically bounded by roads, rivers or other features. This dataset provides index values for small spatial units within urban areas in Zimbabwe. Measures of population density were calculated from high-resolution gridded population datasets from WorldPop, and the space available around buildings was calculated using building footprint polygons derived from satellite imagery (Ecopia.AI and Maxar Technologies. 2020). These data were produced by the WorldPop Research Group at the University of Southampton. This work was part of the GRID3 project with funding from the Bill and Melinda Gates Foundation and the United Kingdom’s Department for International Development. Project partners included the United Nations Population Fund (UNFPA), Center for International Earth Science Information Network (CIESIN) in the Earth Institute at Columbia University, and the Flowminder Foundation.

This dataset consists of land cover maps for the district of Murewa in Zimbabwe, for the year 2023. The dataset was created as part of the RAIZ (Resilience building through agroecological intensification in Zimbabwe) project FOOD/2021/424- 933 (https://raiz.org.zw/), founded by the European Union. The land cover maps were generated using PlanetScope mosaics provided by the NICFI Program. Monthly mosaics, covering the period from June 2022 to June 2023, are provided at a 5m spatial resolution. A deep learning model based on a 3D convolutional neural network (CNN) was used to process the satellite images. The images of the different dates are concatenated, and patches of 17x17 pixels are used as input, with the model predicting the class of the central pixel. This approach enables the model to learn spatial patterns in the images as well as temporal information. The inputs to the model included all available bands (blue, green, red, and near infrared) as well as the Normalized Difference Vegetation Index (NDVI). Land Use Classes: • Dense Woodlands (0): Areas with a dense canopy where grass is not visible from above. Trees are typically over 2 meters tall. • Cropland (1): Land used for agricultural purposes, including fields and farms for crop cultivation. • Grassland (2): Areas dominated by grass, with occasional shrubs and trees covering less than 20% of the area. These are primarily open grasslands. • Open Woodlands (3): Savanna woodlands with a sparse canopy of trees (over 2 meters tall) and scattered shrubs. • Mineral Soils (4): Rocky areas, including granitic mountains and hills, locally referred to as "Dwalas." • Built-up Surface (5): Urbanized areas that include buildings, roads, and other man-made structures. • Bare soil (6): Areas characterized by exposed earth with minimal to no vegetation • Grassland Vlei (7): Wet grassland areas, typically found in low-lying or floodplain regions, where water accumulates seasonally. • Water (8): Bodies of water such as rivers, lakes, and ponds.

The data includes the path data of tropical cyclone "iday" in the southern hemisphere in March 2019, and the data of flood affected area in southern Africa caused by it. It is an important data source supplement for major global tropical cyclone disasters in 2019. The track data of the tropical cyclone is collected from the monitoring data of the National Satellite Meteorological Center, and the longitude and latitude coordinates are obtained by using ArcGIS software; the flooded range data of the southern Africa flood is extracted by the Institute of remote sensing of the Chinese Academy of Sciences Based on the high-resolution three satellite image. The data can be used for the path analysis, affected situation analysis and disaster damage assessment of tropical cyclone "Yidai".

The Global Annual PM2.5 Grids from MODIS, MISR and SeaWiFS Aerosol Optical Depth (AOD) data sets represent a series of three-year running mean grids (1998-2012) of fine particulate matter (solid particles and liquid droplets) that were derived from a combination of MODIS (Moderate Resolution Imaging Spectroradiometer), MISR (Multi-angle Imaging SpectroRadiometer) and SeaWIFS (Sea-Viewing Wide Field-of-View Sensor) AOD satellite retrievals. Together the grids provide a continuous surface of concentrations in micrograms per cubic meter of particulate matter 2.5 micrometers or smaller (PM2.5) for health and environmental research. For each satellite-derived PM2.5 source, the total column retrievals of AOD were converted to near-ground PM2.5 levels using the GEOS-Chem chemical transport model to represent local relationships between AOD and PM2.5. A global decadal (2001-2010) mean PM2.5 concentration grid was also produced. The raster grids have a grid cell resolution of 6 arc-minutes (0.1 degree or approximately 10 km at the equator) and cover the global land surface from 70 degrees north to 55 degrees south. Compared to SEDAC’s earlier data set, Global Annual Average PM2.5 Grids from MODIS and MISR Aerosol Optical Depth (AOD), v1 (2001-2010), this data set provides higher accuracy; longer temporal range; higher resolution (0.1 x 0.1 degrees); and time varying AOD to PM2.5 relationships, necessary for appropriate representation of trends. This data was published in 2015.Source: CIESIN