Chapter 6 and 7 consider methods used to explore spatial data. This chapter covers some preliminary ground by defining in section 5.1 exploratory spatial data analysis (ESDA) with reference to exploratory data analysis (EDA). It then considers in section 5.2 some conceptual models for describing spatial variation. Conceptual models are identified because they help to characterize spatial variation in a non-formal way and so help to clarify what the analyst might look out for (in terms of spatial variation) in undertaking ESDA. Formal, statistical, models are not needed for exploratory spatial data analysis, although some of the exploratory forms of analysis discussed in chapter 7 that involve hypothesis testing need to specify a model for spatial randomness. These are discussed as they are needed. The discussion of statistical models for more general forms of spatial variation is left until chapter 9.

EDA and ESDA

Good (1983) defines exploratory data analysis (EDA) to be a collection of techniques for summarizing data properties (descriptive statistics) but also for detecting patterns in data, identifying unusual or interesting features in data, detecting data errors, distinguishing accidental from important features in a data set, formulating hypotheses from data. EDA techniques may also be used to examine model results, provide evidence on whether model assumptions are met and whether there are influential data effects in model fits.