Trends in quantitative methods III: stressing the visual

The previous two reports on current issues in quantitative geography considered new forms of spatial analysis emanating from 1) an increasing realization that relationships might vary across space and the concomitant recognition that the traditional global forms of analysis are unsuited to investigating such variation (Fotheringham, 1997); and 2) the continued rapid advances in computational power from which new geocomputationally intensive forms of analysis are being developed (Fotheringham, 1998). This third report focuses on visualization as an aid to the understanding of spatial patterns and processes. Some good general discussions of visualization of spatial data can be found in Hearnshaw and Unwin (1994) and the special issue of Environment and Planning B: Planing and Design on ‘Visualisation in urban analysis and design’ (1992). Taylor (1991) states that visualization is a field of computer graphics which addresses both analytical and communication issues in visual representation whilst Cleveland (1993: 1) states that ‘Visualization is critical to data analysis. It provides a front line of attack, revealing intricate structure in data that cannot be absorbed in any other way’. In Cleveland’s view, visualization can have the immediate effect of allowing us to draw a conclusion about a relationship or about some issue of model performance or data quality although sometimes it is not enough and probabilistic inference is then needed to (1993: 12) ‘help calibrate the uncertainty of a less certain issue’. While these definitions make it fairly clear that visualization involves the production of new ways of looking at data to create knowledge, DiBiase (1990) usefully draws an important distinction between private visual thinking or visualization in which the emphasis is on the exploration of data and the construction of knowledge and public visual communication or traditional cartography in which the emphasis is on presentation and the dissemination of knowledge. This places visualization very much in the realm of statistical analysis with the challenge for Progress in Human Geography 23,4 (1999) pp. 597–606