“Classification is, perhaps, the basic procedure by which we impose some sort of order and coherence upon the vast inflow of information from the real world.” Harvey (1969)
Harvey provides a very useful background to the field of classification, especially in a spatial context, but for the basics of classification within a GIS context Mitchell (1999, pp 46-55) and Longley et al. (2005, Ch.12) are probably more useful starting points. These latter works address the question of how to classify and map attribute data associated with polygonal regions or provided as continuous fields. The manuals and online help facilities of the main GIS and image processing packages are also a good source of material for deciding which scheme might be the most appropriate for the dataset you have. Most of these schemes provide a range of standard alternatives for classifying univariate attributes, for example mapping population density by census tract. The attributes selected may be transformed in advance, for example: by producing spatially intensive variables such as densities rather than population totals; and/or by mathematically transforming data, e.g. by classifying log transformed data rather than the original dataset; and/or by normalising datasets (as discussed in Section 4.3.3). Alternatively the classification procedure selected may impose some form of normalisation on the untransformed data (e.g. standard deviation based classification).
Classification needs to be seen in terms of purpose as well as method. However, GIS software places almost no constraints upon a user’s selection of classification method, so it is the responsibility of the analyst to choose appropriate procedures and settings that are fit for purpose. Classification is by no means a peculiarly spatial problem, and most methods apply to almost any kind of data, i.e. there is rarely a specific spatial aspect, such as a contiguity condition, applied. Such facilities are more likely to be found in the spatial autocorrelation and pattern analysis features of GIS toolsets.
