Our objective is to measure the Global Peace Index (GPI) at a monthly frequency, as compared to the official annual GPI score, to have more frequent updates of peacefulness. To tackle this task, we apply machine learning techniques, exploiting information extracted from a new digital data source called Global Data on Events, Location, and Tone (GDELT). We also apply the SHAP (SHapley Additive exPlanation) methodology to identify the most important variables that contribute to the estimation of the index. Through the important variables analysis we highlight each country's profile and we identify and explain the events that drive the errors in the predictions. Our data, machine learning and SHAP code and be found in the following link: https://github.com/VickyVouk/GDELT_GPI_SHAP_project