Geospatial Temporal Crime Prediction Using Convolution and LSTM Neural Networks: Enhancing the Las Vegas Cardiff Model

According to the Department of Justice, more than half of violent crimes go unreported to law enforcement in the United States (Kollar et al., 2018). This data gap reduces the opportunity to implement proven solutions in the areas with the greatest need. In 1996, Dr. Shepherd developed the Cardiff Model with the aim of bringing together hospitals, law enforcement, and community leaders through the sharing of data. We partnered with ongoing efforts to implement the Cardiff Model in Las Vegas, Nevada. Our goal was to provide a geospatial temporal model that can predict the next 30 days of crime. By utilizing the Metropolitan Police Department's (LVMPD) violent crime database, we were able to use a combination of long short-term memory (LSTM) and convolutional neural network (CNN) models to predict where and when violent crimes are likely to occur. Our total crime LSTM model produced an RMSE of 8.621 over a 30-day horizon. When incorporating the spatial component, our CNN and LSTM model produces an MSE of 0.0009 over the same horizon. These findings show that with sufficient latitude and longitude tracked violent crime data, we’re able to accurately produce predictive heat maps. This establishes a framework to expand on the current process which develops heat maps aggregated over historical time periods. By adding to the existing drug overdose heat maps built by Grard et al. (2023), we hope to provide local leadership with the necessary tools to achieve similar reductions in violent crimes seen in Cardiff Projects across the globe.