GLOPOP-S

- In Version 5, we compressed all synthetic population files to make downloading the full dataset easier. - We corrected a minor error that affected a subset of the synthetic population files, so we recommend using only Version 5 of the data. - Python and R code to read the data are available on github: https://github.com/VU-IVM/GLOPOP-S - Please refer to the paper (see paper in Nature Scientific Data https://rdcu.be/dWfKJ) and dataset (doi: https://doi.org/10.7910/DVN/KJC3RH) when using GLOPOP-S. About GLOPOP-S: We present GLOPOP-S, the first global synthetic population database with 1,995,341,998 households and 7,336,281,143 individuals for the year 2015. The households and individuals in GLOPOP-S have the following attributes: age, education, gender, relationship to household head, economic situation, settlement type (urban/rural), household size and household type. To generate each country's synthetic population, we used microdata from the Luxembourg Income Study (LIS) and Demographic and Health Surveys (DHS). Using the iterative proportional updating algorithm, we fit national survey data to regional statistics to account for socio-economic and demographic differences across regions within a country. The regions in the synthetic population correspond to the Global Data Lab (GDL) subnational regions. GLOPOP-S can be used in simulation models, such as (large-scale) agent-based models, where individuals or households are identified as agents that make decisions given their attributes and their environmental or geographical context.