Source: World Bank
Our social science team identified dimensions of the natural, built, and social worlds, and the online language around them, to help collect data for three proof-of-concept models that predict violence in the Democratic Republic of the Congo, population change in the Horn of Africa, and fluctuations in crime levels in small developing island nations. Today, our AI models that study drivers of social risk are already informing the design and implementation of World Bank policy and operations in Africa (and other regions).and the subsequent analysis as well.
in the Democratic Republic of Congo
Focusing on three historically conflict-plagued eastern DRC provinces: North Kivu, South Kivu, and Ituri, the first model predicted changes in the number of conflict events and identified the phenomena most associated with change among thousands of model variables. This included sentiment on sensitive topics such as land, mines, identity and governance, the size of official government reserves, the price of sugar, and copper export volumes.
The model informs country analyzes including the World Bank’s Risk and Resilience Assessments, including the nature, scale, location, sequence, and timing of World Bank-supported activities, such as the Stabilization and Reconstruction Project in the Eastern Democratic Republic of the Congo (STAR Est), which envisions risk financing mechanisms to respond to observed or projected levels of conflict.
in the horn of africa
In the data-poor Horn of Africa, a border region where Ethiopia, Kenya, and Somalia meet, our team produced five years of monthly data on built structures in 56 towns and cities using satellite imagery as a proxy for population change. The second model predicts changes in this surrogate data. Figure 2 shows the importance of conflict events, economic and environmental factors, and societal perceptions of them, as influences on population change. This data informed the design of the DRIVE (Derisking, Inclusion and Valorization of Pastoralist Economies in the Horn of Africa) project, which is currently exploring model adaptations to predict livestock supply chain elements that influence agropastoralist vulnerability. The model will also help design a displacement risk prediction model in the second phase of the Ethiopian Development Response Project to Displacement Impacts.
Figure 2. Relationship between population size change and factors identified in the model in border regions of the Horn of Africa.
