ABSTRACT
In the previous chapter, the focus was on the analysis of spatial variation in single variables, with an emphasis on gridded data. In this chapter, the concern is with spatial variation in single variables represented using areas or points. In Section 4.1, the topic of locally estimated summary statistics is introduced, and in Section 4.2 geographically weighted summary statistics are outlined. These sections build on parts of the previous chapter which were concerned specifically with gridded data. Section 4.3 discusses the analysis of spatial autocorrelation, primarily with reference to areal data. In Section 4.4, some local measures of spatial autocorrelation are detailed. Section 4.5 briefly considers the analysis of spatial association and categorical data. The chapter goes on to review some other issues such as other analytical approaches with connections to the themes explored in the main part of the chapter. The main focus of the following section is on variables available on areal
units of varying size and shape. A large proportion of the work conducted using areal data is in human geography and related disciplines. As such, many of the examples cited below are concerned with socioeconomic data, but some applications in physical geography are suggested. Selected approaches are illustrated using values in a remotely-sensed image that were aggregated over areas represented by vector polygons.
