Welfare estimations from imagery. A test of domain experts ability to rate poverty from visual inspection of satellite imagery

TL;DR

This study evaluates how well domain experts can estimate poverty levels from satellite imagery and identifies visual features that correlate with wealth, aiming to improve machine learning models for poverty prediction.

Contribution

It demonstrates the feasibility of using visual indicators from satellite images for welfare estimation and highlights key features that predict wealth, advancing explainable AI in this domain.

Findings

Experts' ratings correlated with DHS wealth rankings.

Modern roofs and wider roads indicate higher wealth.

Poor road coverage and greenery correlate with lower wealth.

Abstract

The present study uses domain experts to estimate welfare levels and indicators from high-resolution satellite imagery. We use the wealth quintiles from the 2015 Tanzania DHS dataset as ground truth data. We analyse the performance of the visual estimation of relative wealth at the cluster level and compare these with wealth rankings from the DHS survey of 2015 for that country using correlations, ordinal regressions and multinomial logistic regressions. Of the 608 clusters, 115 received the same ratings from human experts and the independent DHS rankings. For 59 percent of the clusters, experts ratings were slightly lower. On the one hand, significant positive predictors of wealth are the presence of modern roofs and wider roads. For instance, the log odds of receiving a rating in a higher quintile on the wealth rankings is 0.917 points higher on average for clusters with buildings with slate or tile roofing compared to those without. On the other hand, significant negative predictors included poor road coverage, low to medium greenery coverage, and low to medium building density. Other key predictors from the multinomial regression model include settlement structure and farm sizes. These findings are significant to the extent that these correlates of wealth and poverty are visually readable from satellite imagery and can be used to train machine learning models in poverty predictions. Using these features for training will contribute to more transparent ML models and, consequently, explainable AI.