A place in history: a guide to using GIS in historical research

CHAPTER 7: SPATIAL ANALYSES OF STATISTICAL DATA IN GIS

7.1 Introduction

Visualisation helps us to understand spatial patterns but often further investigation is needed to describe or explain spatial patterns. When using quantitative data this frequently means that researchers will want to perform some form of statistical analysis. When using GIS, an analysis should not just use conventional statistical techniques which only focus on the attribute data, but should also incorporate the spatial component of the data. Bringing the two together is known as spatial analysis, or geographical data analysis (GDA).

Caution is required when using GIS-based analysis of spatial data, and lessons can be learned from the experience of GIS research in human geography. Early enthusiasm for a GIS-based approach led protagonists to claim that GIS provided a cohesive and scientific framework that could re-unite geography as a discipline. This was presented within a highly quantitative framework (see, for example, Openshaw 1991). Not surprisingly, this led to a backlash that particularly focused on the overtly quantitative and scientific approaches that GIS research was taking at this time. This, it was argued, marked a return to "the very worst sort of positivism" (Taylor 1990, 211). It was also argued that as an academic sub-discipline, GIS lacked a strong epistemology and any treatment of ethical, economic or political issues (see, in particular, Pickles 1995).

As time has gone on, some of the more extreme claims about what GIS may achieve have been moderated, while some of the criticisms of it are being addressed. In particular, it has become clear that GIS has more to offer than simply being a quantitative, positivist, number-crunching operation and, as will be discussed in Chapter 8, it has a potentially significant role to play in qualitative research as well as quantitative. In addition, through Geographic Information Science (GISc, see Chapter 1), the role of quantitative GIS-based work as an academic sub-discipline is being increasingly well defined

While there has been much enthusiasm for including statistical functionality as part of the GIS toolbox among researchers, this has not been mirrored among the GIS software vendors. Many software packages claim to have analytic functionality but this usually refers to the inclusion of overlay, buffering and the other forms of spatial manipulation described in Chapter 4, rather than the more statistical techniques dealt with in this chapter. This has advantages as well as disadvantages. In particular, it encourages the researcher to determine for themselves what types of techniques are relevant to their data, and to devise methods of performing the analysis either within the GIS software or outside it. Effectively this means that at present we are forced to think about what techniques are appropriate to our data rather than simply being spoon-fed with vendor-provided 'solutions' that may or may not be appropriate. While this may lead to cumbersome computing, it is hoped that it should lead to more appropriate analyses.