Applying Artificial Intelligence on Satellite Imagery to Compile Granular Poverty Statistics

Abstract

This study outlines a computational framework to enhance the spatial granularity of government-published poverty estimates, citing data from the Philippines and Thailand. Computer vision techniques were applied on publicly available medium resolution satellite imagery, household surveys, and census data from the two countries. The results suggest that even using publicly accessible satellite imagery, predictions generally aligned with the distributional structure of government-published poverty estimates after calibration. The study further examines the robustness of the resulting estimates to user-specified algorithmic parameters and model specifications.