The 'ViolenceDynamics' grant proposal aims to tackle key challenges in conflict prediction and prevention, including unexpected violence outbreaks, lack of understanding of intervention effectiveness, and the absence of a dynamic forecasting framework. The project integrates machine learning with macroeconomic modeling to enhance decision-making. It focuses on predicting local violence using newspaper text analysis and assessing the impact of policies on conflict intensity, migration, gender inequality, and economic damage. Findings will be publicly available on conflictforecast.org, offering monthly predictions, historical policy data, and estimated costs of inaction. The goal is to improve conflict prediction, evaluate policy effectiveness, and support decision-makers in mitigating conflicts' global impact.