Mass atrocities place thousands of humans in dire situations every year—and cost governments millions of dollars.

A non-governmental organization (NGO) wanted to save lives and reduce economic burdens by better understanding and predicting the conditions that indicate a risk of atrocities. Subjective analysis models had given the organization a start towards their goal, but they wanted an automated algorithmic solution that improves the accuracy of predictions.

The NGO believed that newly validated data sets on national and sub-national violence could help them out-predict the existing subjective analysis models, but there was no proof than an answer was present in the data sources.

With an Analytics Starter Pack, Topcoder helped the NGO leverage the new data sets to develop multiple predictive models. Each model identifies the risk factors that make communities more or less likely to experience acts of mass violence, and the top solutions demonstrate a significant event-prediction correlation—proving the NGO's assumption that an improved statistical model could be created.

Topcoder delivered five winning analytics solutions in just three weeks, giving the NGO multiple innovative approaches that can be built upon to better predict and prevent atrocities around the world.

Data Scientists
93
Submissions
618
Winning Solutions
5
Duration
3 Weeks