Seasonal characteristics of crime: an empirical investigation of the temporal fluctuation of the different types of crime in London

Abstract

Abstract Most types of crimes show seasonal fluctuations but the difference and similarity of the periodicity between different crimes are understudied. Interpreting the seasonality of different crime types and formulating clusters of crimes that share similar seasonal characteristics would help identify the common underlying factors and revise the patterns of patrolling and monitoring to enable sustained management of the control strategies. This study proposes a new methodological framework for measuring similarities and differences in the timing of peaks and troughs, as well as the waveforms of different crimes. The method combines a Poisson state-space model with cluster analysis and multi-dimensional scaling. A case study using twelve types of crimes in London (2013–2020) demonstrated that the amplitude of the seasonal fluctuation identified by this method explained 95.2% of the similarity in their waveforms, while the timing of the peaks covered 87.5% of the variance in their seasonal fluctuation. The high predictability of the seasonal patterns of crimes as well as the stable categorisation of crimes with similar seasonal characteristics enable sustainable and measured planning of police resource allocation and, thereby, facilitates a more efficient management of the urban environment.

Keywords

• Cluster analysis
• Crime seasonality
• Multi-dimensional scaling
• Trend analysis
• Urban sustainability