Deep Learning for Crime Forecasting: The Role of Mobility at Fine-grained Spatiotemporal Scales

• In Event: Advances in Quantitative Methods Panel 1: Machine Learning and AI for Crime Analysis

Abstract

Objectives: To develop a deep learning framework to evaluate if and how incorporating micro-level mobility features, alongside historical crime and sociodemographic data, enhances predictive performance in crime forecasting at fine-grained spatial and temporal resolutions.

Methods: We advance the literature on computational methods and crime forecasting by focusing on four U.S. cities (i.e., Baltimore, Chicago, Los Angeles, and Philadelphia). We employ crime incident data obtained from each city's police department, combined with sociodemographic data from the American Community Survey and human mobility data from Advan, collected from 2019 to 2023. This data is aggregated into grids with equally sized cells of 0.077 sq. miles and used to train our deep learning forecasting model that predicts crime occurrences 12 hours ahead using 14-day and 2-day input sequences.

Results: Incorporating mobility features improves predictive performance, with our deep learning model achieving the highest recall in all four cities, outperforming alternative methods. Notably, longer input sequences enhance predictions for violent crimes, while shorter sequences are more effective for property crimes, particularly when augmented with mobility data.

Conclusion: These findings underscore the importance of integrating diverse data sources for spatiotemporal crime forecasting. They also highlight the advantages (and limits) of deep learning when dealing with fine-grained spatial and temporal scales.

Authors

• Ariadna Albors Zumel, University of Trento

• Michele Tizzoni, University of Trento

• Gian Maria Campedelli, University of Trento