USAID/HU Tech Challenge for Atrocity Forecasting Uses GDELT

The United States Agency for International Development (USAID) and Humanity United Model Challenge, the final round of the Tech Challenge for Atrocity Prevention, "asked innovative problem solvers to create algorithmic models that can help forecast when and where mass atrocities are likely to occur." Nearly 100 contributors submitted 618 algorithms to the competition, with the winning submission being "especially successful at predicting atrocities in regions with limited or no past history of mass violence, taking into account 23 diverse geographic, socio-political and historical violence factors within each region." The Political Instability Task Force (PITF) data on worldwide atrocities was the gold event set that participants had to forecast, while historic PITF data and GDELT were the two input datasets. Read the press release or see the list of winners of the competition. Read this article for more details on the winning algorithms.