Transfer learning improves socioeconomic data for displaced populations
Transfer Learning for Socioeconomic Estimation in Forced-Displacement Settings
Machine LearningArtificial Intelligence
Summary
Collecting detailed household data about people forced to move is slow and costly, and things change quickly in these situations. The authors improved a computer program that uses satellite images and machine learning by training it on lots of data from African countries. They adapted it to better estimate living conditions in refugee camps and nearby areas in South Sudan, Cameroon, and Zambia. This helps fill gaps between big surveys, giving more timely and detailed information to help humanitarian aid reach the right places.
What this means in practice
- •For humanitarian organizations: Estimate changes in living conditions of displaced populations regularly using satellite data to better target aid and prioritize field visits.
- •For government planning teams: Monitor socioeconomic trends in areas affected by conflict using updated model-based estimates between national household surveys.
Authors
Steven Ndung'u, Adel Daoud, Ismael Yacoubou Djima, Hai-Anh H. Dang, Patrick Michael Brock
Abstract
Progress in inclusive household surveys has strengthened socioeconomic evidence for forcibly displaced populations, providing indispensable benchmarks on living conditions and welfare. However, these surveys remain resource-intensive and periodic, while conditions can change between rounds, particularly in settings affected by fragility, conflict, and violence. More frequently updated, spatially granular complementary evidence is therefore needed to identify where socioeconomic conditions may be changing between survey rounds and to inform operational prioritization. Earth observation and machine learning offer a scalable source of spatially explicit socioeconomic information. However, tools developed for general populations have not been systematically adapted and evaluated in forced displacement settings, where living conditions, settlement patterns, and displacement impacts may differ substantially. We address this gap by adapting a multimodal spatiotemporal vision transformer, pretrained on Demographic and Health Survey data from approximately 1.2 million households across 36 African countries, to forced displacement and host community settings in South Sudan, Cameroon, and Zambia. We develop and evaluate the updated, adapted model using socioeconomic indices derived from UNHCR FDS and RMS data. Our results show that satellite-derived geospatial covariates explain up to 66% of the variation in socioeconomic outcomes in camp-intersecting grids, with a mean absolute error (MAE) of 4.37 index points, and 41% in non-camp-intersecting areas, with an MAE of 5.41. The framework complements and adds value to periodic household surveys by filling critical spatial and temporal data gaps with regularly updated, model-based socioeconomic estimates. These estimates sustain insight between survey rounds and support timely humanitarian prioritization and field verification.