Part II — Spatial Data Analysis

Overview

Now that we understand geographic processes and the data that measures them, we will introduce exploratory spatial data analysis (ESDA). ESDA augments Tukey’s exploratory data analysis, and involves a large collection of techniques used to “orient yourself” (find structure) inside of your dataset. For geographical problems, this often involves understanding whether our data displays a geographical pattern. We cover such topics in this part.

First, in Chapter 5, we discuss the workhorse of statistical visualization for geographic data: choropleths. In Chapter 6, we introduce spatial autocorrelation, the concept that formally connects geographical and statistical similarity. This allows us to characterize the “strength” of a geographical pattern and is at the intellectual core of many explicitly spatial techniques. All patterns have exceptions, however, and Chapter 7 will present local methods that can detect observations that are unlike (or too like) their neighbors. To wrap up, Chapter 8 discusses methods for visualising, characterising and analysing points, the raw locational data. Taken altogether, this part provides methods to explore most of the fundamental questions involved in geographical analysis: whatever the nature of my data, is there a geographical pattern, and are there places where this pattern does not hold?