Using publicly available satellite imagery and deep learning to understand economic well-being in Africa

By Christopher Yeh, Anthony Perez, Anne Driscoll, George Azzari, Zhongyi Tang, David B. Lobell, Stefano Ermon, Marshall Burke. Nature Communications, 2020

460 citations recorded by the source corpus.

Accurate and comprehensive measurements of economic well-being are fundamental inputs into both research and policy, but such measures are unavailable at a local level in many parts of the world. Here we train deep learning models to predict survey-based estimates of asset wealth across ~ 20,000 African villages from publicly-available multispectral satellite imagery. Models can explain 70% of the variation in ground-measured village wealth in countries where the model was not trained, outperforming previous benchmarks from high-resolution imagery, and comparison with independent wealth measurements from censuses suggests that errors in satellite estimates are comparable to errors in existing ground data. Satellite-based estimates can also explain up to 50% of the variation in district-aggregated changes in wealth over time, with daytime imagery particularly useful in this task. We demonstrate the utility of satellite-based estimates for research and policy, and demonstrate their scalability by creating a wealth map for Africa’s most populous country. It is generally difficult to scale derived estimates and understand the accuracy across locations for passively-collected data sources, such as mobile phones and satellite imagery. Here the authors show that their trained deep learning models are able to explain 70% of the variation in ground-measured village wealth in held-out countries, outperforming previous benchmarks from high-resolution imagery with errors comparable to that of existing ground data.

Subfields: Global and Planetary Change

View Paper: https://www.semanticscholar.org/paper/83bd44a487ea02e19a27e9d77cd736dd4f5bcc00

Citations in the corpus

• Generalizing to Unseen Domains: A Survey on Domain Generalization (2021, unknown): “Dataset #Domain #Class #Sample Description Reference Office-Caltech 4 10 2,533 Caltech, Amazon, Webcam, DSLR [224] Office-31 3 31 4,110 Amazon, Webcam, DSLR [224] PACS 4 7 9,991 Art, Cartoon, Photos, Sketches [2] VLCS 4 5 10,729 Caltech101, LabelMe, SUN09, VOC2007 [225] Office-Home 4 65 15,588 Art, ”
• Using satellite imagery to understand and promote sustainable development (2019, unknown): “To further explore this ability to overcome noise, we used data from an earlier study of African asset wealth (28) to explore the influence onmodel performance of three types of errors common in publicly available training data: (i) random noise (“jitter”) purposely added to village geocoordinates t”
• In-N-Out: Pre-Training and Self-Training using Auxiliary Information for Out-of-Distribution Robustness (2020, unknown): “In remote sensing, satellite imagery is paired with aux-inputs to predict the desired output [49, 54].”
• WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, unknown): “PovertyMap-wilds is based on a dataset collected by Yeh et al. (2020), which assembles satellite imagery and survey data at 19,669 villages from 23 African countries between 2009 and 2016 (Figure 4).”