The ViEWS project will start up in January 2017 with funding from the ERC Advanced Grant scheme. It will develop, test, and iteratively improve a pilot Violence Early-Warning System (ViEWS) that is rigorous, data-based, and publicly available to researchers and the international community. ViEWS will provide early warnings for the three forms of political violence recorded by the UCDP: armed conflict involving states and rebel groups, armed conflict between non-state actors, violence against civilians, and forced population displacement, and apply these to specific actors, sub-national geographical units, and countries.
The figure above illustrates forecasts at the geographical unit level. The black dots represent UCDP conflict events recorded for January 2012. The colors represent forecasted risk of such events based on data up to December 2011. The model displayed here is very simple and only takes the history of past conflict events into account. Core tasks of the ViEWS project will be to add detailed geographical information, link to the other levels of analysis in the system, and iteratively improve the forecasts by means of out-of-sample evaluation and Bayesian Model Averaging.