This paper highlights the role and importance of street patterns in crime simulation and analysis. First, many crimes follow street patterns in reality. However, conventional statistical approaches to crime clustering generate misleading clusters that deviate from the underlying street patterns. An example is provided to demonstrate that taking street patterns into consideration during the clustering process can alleviate this problem. Second, most statistical analyses of crimes are based on a set of arbitrarily chosen area units such as zipcode areas, census tracts and city neighborhood areas. The results of such analyses are inherently biased due to the problem of the modifiable area unit problem (MAUP), i.e., varying area aggregations lead to different results. This paper demonstrates that the consideration of street patterns during the process of area aggregation can help reduce MAUP bias. Finally, a simulation of street robberies illustrates how street patterns influence the spatial learning and behavior of offenders and victims and the interaction between them. With these examples, the author hopes more emphasis be place on spatial patterns such as street patterns in crime simulation and analysis.
Audio (MP3 File, Podcast Ready)
Copyright. All Rights Reserved.