crime hotspot analysis techniques use straight-line (Euclidean) distance and assume that space is continuous, homogeneous and uniform in all directions. In other words, these techniques assume that crime can happen anywhere. But many crime events are constrained by a one-dimensional subset of this space, network space: • Street robbery • Bus-related crime • IED attacks • Perhaps any crime geocoded to a property address
to network constrained events ….. • Ignores the layout of urban space • Underestimates actual travel distance • Can produce spurious evidence of clustering when applied to network constrained data like road collisions (Yamada and Thill 2004) and vehicle thefts (Lu & Chen 2007)
84 Driver Incident Reports of Disturbance during 2013 on a single bus route run. OS MasterMap Integrated Transport Layer. ! ! Planar K-function using spatstat package in R. 1m distance interval. No edge correction. 99 simulations. Network K-function in GeoDa Net. 1m distance interval. Edge correction unavailable. 99 simulations.
present in both outputs of the K function. Therefore, clustering is not a function of incidents being constrained by a bus route. ! So what is responsible?……
analysis techniques that use shortest path distance are recommended for network constrained crime events to reduce the detection of potentially spurious clustering patterns. ! Where data is highly constrained (e.g. bus-related crime) it is necessary to adopt network spatial methods. !
& Chen, X. (2007). On the false alarm of planar K-function when analyzing urban crime distributed along streets. Social Science Research, 36 (2) 611-632. Okabe, A. & Yamada, I. (2001). The K-function method on a network and its computational implementation. Geographical Analysis, 33, 271–290. Okabe, A. & Sugihara, K. (2012). Spatial Analysis Along Networks. Chichester, West Sussex: John Wiley & Sons. Tompson, L, Partridge, H, & Shepherd, N. (2009). Hot Routes: Developing a New Technique for the Spatial Analysis of Crime. Crime Mapping: A Journal of Research and Practice, 1 (1) 77-96. Yamada, I. & Thill, J-C. (2004). Comparison of planar and network K-functions in traffic accident analysis. Journal of Transport Geography, 12, 149-158. !