Conflict Cartographer: Conflict expert prediction in Africa

Led by Siri Aas Rustad

Oct 2020 – Dec 2023

​Accurately predicting where a conflict will escalate or take place can provide invaluable insights for policymakers, peacekeepers and other important stakeholders.

Accurately predicting where a conflict will escalate or take place can provide invaluable insights for policymakers, peacekeepers and other important stakeholders. This knowledge can be used to implement preventive measures in vulnerable areas, potentially alleviating human suffering. This has been the primary aim of the increased efforts to predict conflict that has evolved over the past few years. Most forecasts use advanced statistical models and large amounts of data. The results from these predictions indicate that they are able to produce quite accurate predictions of armed conflict, especially when it comes to forecasting escalation from low- to high levels of violence.

The models, however, have some limitations. They are dependent on new data, which for certain important predictors are not updated frequently enough. Secondly, they are not so good at predicting escalation from a state of no violent conflict to minor violent conflict. Lastly, they struggle with narrowing down the subnational area where conflict will happen. To address these issues, the Conflict Cartographer project aims to complement existing quantitative prediction models using qualitative expert opinion surveys. The project intends to use local- and expert knowledge to investigate in what areas conflict is likely to take place, while also investigating whether conflict will erupt. The project thus aims to fill some important gaps in conflict prediction, hopefully improving the ability to target vulnerable areas on a sub-national and local level, in addition to forecasting conflict onset.

Expert opinion surveys are a great alternative when the matter at hand is complex and difficult to assess in other ways. Within conflict research there are very few expert opinion surveys, and this project builds on a conflict expert opinion survey that we conducted through EU-Listco, a Horizon2020 project. Due to limitations within the EU-Listco project, only five countries were investigated. In this new project, we focus on 35 countries in Africa that have experienced conflict within the past five years.

In order to complement existing quantitative prediction models, we invite participants, through an online app, to draw on a map where they believe violent conflict is likely to take place the next three months. Given that they believe conflict will occur, they will be asked what the intensity will be, in terms of people killed. We will then compare the results from the expert surveys with other prediction models, and compare these results against the actual conflict that occurred in the areas. This way, we are able to assess the predictability of our models, and hopefully improve forecasting.

To get as accurate predictions as possible, we hope to attract as many people as possible that have any knowledge on conflict in Africa to share their predictions with us. The exercise is open for all. To participate go to

Questions about the project can be directed to

This project is part of the Conflict Trends project, which is funded by the Norwegian Ministry of Foreign Affairs.

An error has occurred. This application may no longer respond until reloaded. An unhandled exception has occurred. See browser dev tools for details. Reload 🗙