University of Chicago data and social scientists have developed a new algorithm that forecasts crime by learning patterns in time and geographic locations from public data on violent and property crimes.
It has demonstrated success at predicting future crimes one week in advance with approximately 90% accuracy.
