Poverty maps help governments and NGOs track socioeconomic trends and allocate resources where they are needed most.
This interactive tool allows users to explore and compare inferred wealth across countries using different machine learning models and multimodal data.
Africa
We estimate wealth in Sierra Leone, Uganda, Liberia, Rwanda, and Gabon, using machine learning models trained on seven freely available data sources, including satellite images, metadata, crowdsourced information, and social media. Wealth is measured using the International Wealth Index (IWI), which ranges from 0 (lowest socioeconomic class) to 100 (highest).
Each mapped location represents multiple households, colored by average wealth. Most features were extracted in 2022 and 2023. Models were trained on survey data from 2016 and 2018-19. The comparison tab shows changes in wealth estimates between these years.
Models
Feature-based: the CatBoost model, is trained with metadata features (i.e., Demographics, Mobility, Population, Nightlight intensity, Antennas, and Infrastructure).
Image-based: the Convolutional Neural Network model with daylight satellite images.
Combined: uses all features.
Weighted: “weighted samples” applied during training to correct for wealth imbalance.
Augmented: “offline augmentation” applied to correct for small samples.
Publication
Lisette Espín-Noboa, János Kertész, and Márton Karsai. 2023. Interpreting wealth distribution via poverty map inference using multimodal data. In Proceedings of the ACM Web Conference 2023 (WWW ’23), April 30-May 4, 2023, Austin, TX, USA. ACM, New York, NY, USA, 12 pages. https://doi.org/10.1145/3543507.3583862
In Guayaquil, Ecuador’s largest city, we infer wealth using supermarket transactions and machine learning. Here, per capita income (in USD) serves as a wealth proxy. Households scoring close to 100 represent the lowest socioeconomic class, while those near 500 belong to the highest.
The map displays wealth predictions for the fourth trimester of 2018, while the comparison tab allows users to compare estimates between the first and fourth trimesters of 2018.
Publication
Cruz, E., Villavicencio, M., Vaca, C. et al. A new approach to estimate neighborhood socioeconomic status using supermarket transactions and GNNs. EPJ Data Sci. 14, 3 (2025). https://doi.org/10.1140/epjds/s13688-024-00515-9
