Crime Mapping and Spatial Data Analysis using R

A key piece of information we explore in environmental criminology is the "exact" location of crimes, typically in the form of street address in which a crime is recorded as having occurred. A key goal of the analysis of this type of mapped point data is to detect patterns, in particular to detect areas where crime locations appear as clustered and reflecting an increased likelihood of occurrence.

In Chapter 1, we generated some maps with point locations representing crimes in Greater Manchester (UK), and we saw how the large volume of events made it difficult to discern any spatial pattern. In previous chapters we saw how we could deal with this via aggregation of the point data to enumeration areas such as census tracts or by virtue of binning. In this chapter we will introduce techniques that are used as an alternative visualisation for point patterns and that are based on the mathematics of density estimation. They are used to produce isarithmic maps; these are the kind of maps you often see in the weather reports displaying temperature. What we are doing is creating an interpolated surface from discrete data points. Or in simpler terms we are "effectively inventing data values for the areas on the map for which you don’t have data or sample points" [155].

Before we get to the detail of how we produce these maps, we will briefly and very generally introduce the field of spatial point pattern analysis. The analysis of discrete locations are a sub-field of spatial statistics. In this chapter we will introduce an R package, called spatstat, that was developed for spatial point pattern analysis and modelling. It was written by Adrian Baddeley and Rolf Turner. There is a webpage and a book [156] dedicated to this package. In our book we are only going to provide you with an introductory practical entry into this field of techniques. If this package is not installed in your machine, make sure you install it before we carry on. In this chapter we will also be using the package crimedata developed by our colleague Matt Ashby, which we introduced in chapter 6 [157].

# Packages for reading data and data carpentry
library(readr)
library(dplyr)

# Packages for handling spatial data and for geospatial carpentry
library(sf)
library(maptools)
library(raster) # used for data represented in a raster model

# Specific packages for spatial point pattern analysis
library(spatstat)

# Libraries for spatial visualisation
library(tmap)
library(ggplot2)
library(leaflet)

# Libraries with relevant data for the chapter
library(crimedata)