Linking Synthetic Populations to Household Geolocations: A Demonstration in Namibia

Dana Thomson-Browne
Lieke Kools, Leiden University
Warren Jochem , University of Southampton

To evaluate gridded population estimates and household survey sample designs, a population census linked to residential locations are needed. Geolocated census microdata data, however, are rarely available and are thus best simulated. In this paper, we simulate a close-to-reality population of individuals nested in households geolocated to realistic building locations using the R simPop package and ArcGIS. Multiple realizations of a geolocated synthetic population are derived from Namibia 2011 census 20% microdata sample, Namibia 2013 Demographic and Health Survey (DHS), and dozens of spatial covariates. We link simulated households to manually generated latitude-longitude coordinates of buildings, and model probability surfaces of distinct household types using Random Forest methods. We simulate five realizations of a synthetic population in Namibia’s Oshikoto region at the level of household, woman, and child. Comparison of variables in the synthetic population with census and DHS data indicate similar distributions and spatial patterns.

See paper

 Presented in Session 3. Population, Development, & the Environment; Data & Methods; Applied Demography