The Violence & Impacts Early-Warning System
A leading open-source platform for forecasting armed conflict and humanitarian crisis
Open Data for Early Warning and Early Action
We provide monthly forecasts on the occurrence, extent, and intensity of state-based armed conflict worldwide, covering 1 to 36 months ahead at a resolution up to 0.5° (55 sq. km at the Equator).
Our open data is available through an interactive forecasting tool, downloadable datasets, and API, empowering users with data-driven insights to stay ahead of emerging crises.
Leverage our platform to:
- Identify regions at heightened risk of armed conflict
- Assess the extent and severity of impending violence
- Estimate conflict-induced humanitarian needs
- Contextualize conflict alerts against baselines and historic data
- Support proactive strategies for conflict mitigation and peacebuilding
Our models are informed by well-maintained and fine-grained data from over a dozen trusted data providers, capturing hundreds of features ranging from conflict history, economic growth, and political stability, to natural resource deposits, natural terrain, and societal vulnerability to natural hazards and climate extremes.
Stay informed. Take action. Strengthen resilience.
Powered by State-of-the-Art Technology
Our platform offers an open, industry-grade MLOps, DevOps, and CI/CD forecasting solution with built-in drift detection, quality assurances, and real-time monitoring that ensures robust, transparent, and reliant service provision for informed decision-making. Key features include:
- Advanced Logging and Monitoring: Real-time logging to support quality assurance and rapid issue resolution.
- Version Control and Traceability: Semantic versioning of all pipeline and model releases to facilitate reproducibility and rollback.
- Branching Strategy for Stability: A clear separation of production and development workflows, featuring robust versioning, hot-fix capabilities, and a staging environment for pre-deployment validation of new models.
- On- & Offline Evaluation: Monthly assessments of model predictions against the most recent conflict data, coupled with regular out-of-sample evaluation on partitioned training and test datasets against the thoroughly vetted annual releases of UCDP GED data, providing crucial insights into model performance for research and model development.
