Fine-grained Population Mapping from Coarse Census Counts and Open Geodata

TL;DR

This paper introduces POMELO, a deep learning model that generates detailed 100m resolution population maps from coarse census data and open geospatial information, even without census counts.

Contribution

POMELO is the first model to accurately produce fine-grained population maps from coarse data and open geodata, capable of generalizing across countries without census counts.

Findings

Achieves 85-89% R2 with disaggregated census data.

Reaches 48-69% R2 when no census data is available.

Effective across multiple sub-Saharan African countries.

Abstract

Fine-grained population maps are needed in several domains, like urban planning, environmental monitoring, public health, and humanitarian operations. Unfortunately, in many countries only aggregate census counts over large spatial units are collected, moreover, these are not always up-to-date. We present POMELO, a deep learning model that employs coarse census counts and open geodata to estimate fine-grained population maps with 100m ground sampling distance. Moreover, the model can also estimate population numbers when no census counts at all are available, by generalizing across countries. In a series of experiments for several countries in sub-Saharan Africa, the maps produced with POMELOare in good agreement with the most detailed available reference counts: disaggregation of coarse census counts reaches R2 values of 85-89%; unconstrained prediction in the absence of any counts reaches 48-69%.