Discovering spatial interaction patterns of near repeat crime by spatial association rules mining.

Area of Science:

• Urban studies
• Criminology
• Data science

Background:

• Crime incidents exhibit complex spatio-temporal dependencies.
• Previous research highlights spatial clustering but overlooks anisotropic interactions and detailed transmission pathways.
• Understanding these dynamics is crucial for effective crime prevention and policing.

Purpose of the Study:

• To develop a new spatial association mining approach for discovering significant crime transmission routes and high-flow regions.
• To analyze anisotropic spatial interaction characteristics in urban crime.
• To explore detailed spatial transmission patterns of crime occurrences.

Main Methods:

• Identification of near-repeat crime pairs based on spatio-temporal proximity.
• Spatial aggregation of crime links onto spatial grids.
• Definition of spatio-temporal interaction measures and development of a spatial association pattern mining approach.
• Analysis of the relationship between transmission patterns and road network structure.

Main Results:

• The proposed approach effectively identifies significant spatial transmission patterns from large-scale crime data.
• High-flow regions and detailed spatial transmission routes of crime are discovered.
• The study reveals insights into the anisotropic nature of spatial crime interactions.

Conclusions:

• The developed method provides a robust framework for analyzing urban crime spatio-temporal dynamics.
• Findings offer valuable guidance for crime pattern analysis and targeted crime prevention efforts.
• The approach can inform data-driven policing strategies by highlighting key transmission pathways.