Crime Mapping and Spatial Data Analysis using R

The focus in this chapter has been to introduce the issue of crime events along a network. There has been increasing recognition in recent years that the spatial existence of many phenomena is constrained by networks. We have discussed some of the issues associated with the storage, handling, and visualisation of this kind of data. Chapter 1 of [Okabe and Sugihara (2012)] provides a fairly accessible introduction to the relevance of point pattern data along a network, whereas Chapter 3 offers a slightly more technical discussion of the computational issues associated with the statistical analysis of points along a network. The first few sections of chapter 17 of [Baddeley et al. (2015)] provide background and very useful detail on the way that spatstat handles and stores point pattern data in a network. [Baddeley et al. (2021)] offers an excellent overview of the challenges of analysing this kind of data and the available methods we have. This overview can also offers an excellent framework to understand the issues of clustering and detection of local clusters along networks to which we will return in the next two chapters.

But spatial networks analysis is a more general area, with various other applications that we could not explore in detail here. [O’Sullivan (2014)] provides an introduction to spatial network analysis, basically a translation of social network metrics and concepts applied to spatial networks. This kind of perspective is particularly helpful within transport, routing, and similar applications. See, for example, the kind of functionality and applications provided by stplanr. We are also witnessing cross-fertilisation between social network analysis and spatial analysis. In epidemiology, for example, we see how this is done for studying the transmission of disease through personal and geographic networks (see, for example, [Emch et al., 2010]). Whereas [Radil et al. (2010)] and [Tita and Radil (2011)] offer some criminological examples (focused on the study of gangs) linking social network analysis and spatial analysis. Along these lines, the package spnet facilitates the rendering of social network data into a geographical space. Some of these applications can be relevant for investigative crime analysis and other criminological uses. Finally, another way in which social networks are being used within the context of environmental criminology is in contributing to the alternative specifications of neighbourhoods that aim to move beyond the traditional use of census geographies (see [Hipp et al., 2012] and [Hipp and Boessen, 2013]). Finally, it is also worth exploring the literature on space syntax and crime, and the various analyses that are looking at the structural characteristics of the street network (such as betweenness) and how these characteristics influence crime ([Davies and Bowers, 2018]; [Kim and Hipp, 2019]).