Working Paper: NBER ID: w27879
Authors: Marshall Burke; Anne Driscoll; David Lobell; Stefano Ermon
Abstract: 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 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 models’ predictive performance. We quantify recent model performance across multiple sustainable development domains, discuss research and policy applications, explore constraints to future progress, and highlight key research directions for the field.
Keywords: Satellite Imagery; Sustainable Development; Machine Learning
JEL Codes: C45; C55; O1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| satellite technology advancement (O30) | predictive model accuracy (C52) |
| improved satellite imagery (Y91) | data accuracy (Y10) |
| availability of quality data (C80) | effectiveness of satellite-based models (C52) |
| high-quality local training data (J24) | model performance (C52) |
| improved training data (J24) | better model performance (C52) |
| satellite data complements ground-based data collection (Q54) | effectiveness of data collection (C80) |