Using satellite imagery to understand and promote sustainable development

Accurate and comprehensive measurements of a range of sustainable development outcomes are fundamental inputs into both research and policy. We synthesize the growing literature that uses satellite imagery to understand these outcomes, with a focus on approaches that combine imagery with machine learning. We quantify the paucity of ground data on key human-related outcomes and the growing abundance and improving resolution (spatial, temporal, and spectral) of satellite imagery. We then review recent machine learning approaches to model-building in the context of scarce and noisy training data, highlighting how this noise often leads to incorrect assessment of model performance. We quantify recent model performance across multiple sustainable development domains, discuss research and policy applications, explore constraints to future progress, and highlight research directions for the field.

Automated summary

Accurate and comprehensive measurements of sustainable development outcomes are essential for research and policy making. The study reviews the use of satellite imagery and machine learning for understanding these outcomes, highlighting the scarcity of ground data and the improving resolution of satellite imagery. It also discusses the challenges of assessing model performance correctly in the context of noisy training data, and explores constraints to future progress and research directions in the field.