Scientific Visualisation and the Exploratory Analysis of Area Data

Abstract

Area data are commonly used in many social science disciplines, but have a number of characteristics which have necessitated the development of specialist tools for analysis and display. In this paper we discuss the application to area data of exploratory analysis methods and visualisation techniques, and in particular the growing number of software packages which provide visualisation methods to support exploratory spatial data analysis (ESDA). In the exploratory analysis of nonspatial data, the data are often considered to consist of rough and smooth elements. If this idea is extended to spatial data, it can be argued that a software system for ESDA should have facilities for exploring the rough and smooth elements of both the spatial and nonspatial elements of the data. Visualisation techniques can assist with this by providing a wide range of graphical displays, in a highly interactive environment. Conceptual frameworks are developed for both ESDA and scientific visualisation, and four software packages that link graphical and cartographical views of the data are reviewed against these requirements. The packages fall into two broad groups. In one, the emphasis is on highly interactive graphics, but with few numerical methods or tools for exploring the spatial properties of the data. In the other, specialist spatial tools are provided, but the graphical tools are less flexible. It is argued that a combination of both approaches would provide a package which more fully satisfied the requirements of ESDA. The need to assess the effectiveness of different visualisation techniques in supporting and encouraging the exploratory analysis of spatial data is noted.