Spatially disaggregated population maps are crucial for accurate impact assessments and planning humanitarian operations, in particular, those focusing on epidemics and public health, food security, conflicts, natural disasters, and a changing climate. Existing gridded population data, however, suffer from limitations driven by the applications of simple dasymetric mapping methods. While data produced using these methods works for certain contexts, we show that these methods do not produce accurate population maps for low-income and vulnerable regions. In this study, we propose a novel method to disaggregate administrative level population data into high-resolution pixel-level population density estimates by measuring cluster level population density using a conditional multiple regression method.
This reopsitory includes:
- the python notebook files which are used to creat the population desntiy estimates
- population density files in GoeTiff format for 140 countries (each file name includes ISO coudes for each country)