Identifying the appropriate spatial resolution for the analysis of crime patterns

Area of Science:

• Spatial analysis
• Geographic information systems (GIS)
• Quantitative criminology

Background:

• Selecting an appropriate scale is crucial for spatial process analysis.
• Smaller geographical units are preferred for human phenomena but face data and quantitative challenges.
• Existing methods struggle to identify optimal aggregation scales for point data.

Purpose of the Study:

• To present a novel approach for estimating the most appropriate scale to aggregate point data into areas.
• To address the challenge of scale selection in spatial analysis.

Main Methods:

• The method involves creating regular grids with decreasing cell sizes (increasing resolution).
• It estimates the similarity between point pattern realisations at each resolution.
• Applied to simulated data and real crime data (Vancouver: residential/commercial burglary, theft from vehicle/bike).

Main Results:

• The study identified appropriate spatial unit sizes for different crime types.
• Results are influenced by data event numbers and spatial clustering, indicating no single universal scale.
• The method provides a means to better estimate appropriate spatial scales for specific analyses.

Conclusions:

• The developed method aids in selecting optimal spatial scales for point data aggregation.
• It offers a flexible approach adaptable to various spatial datasets and analytical needs.
• Accurate scale selection is vital for robust spatial process analysis and understanding geographic patterns.