Linking spatial patterns to processes

This chapter examines the thorny issues underlying many spatial analysis techniques: how we relate the analysis of expressed patterns in geographic data to the underlying processes we are interested in making inferences about. In many fields that rely on quantitative analysis causal inference methods have gained popularity in recent years to overcome limitations of making conclusions from statistical associations in observational data. We examine this trend in the context of geographical analysis, and then consider how the current explosion in production of geographically referenced data changes our ability to learn about underlying spatial processes. We first explore these ideas in terms of well-developed ideas associated with causal theory and thinking. We then examine pattern process linkages through the lens of complexity, as we are often dealing with complex processes driving observable patterns. Finally, we discuss the issues of error and uncertainty in the context of spatial pattern measurement and the generation of spatial processes. A case study into the classic Cholera outbreak in London, 1854 (as mapped by John Snow) is used to illustrate some of the key issues and concepts covered in the chapter. Finally, we conclude with three ideas areas of future exploration (i. Spatial Comparison, ii) Spatial Simulation, and iii) Spatial Theory Development) that postulate how quantitative geography can begin to incorporate causal thinking into analytic methodology and develop tools to facilitate representation and investigation of realistic spatial processes.