points can be used to give an interpretation to the map, but also as a navigation tool in the map based interface as is done in our product catalog map using NL-PCA, which is introduced in the next section. Figure 17.4 shows the NL-PCA configuration of the example product catalog which was made using CATPCA also available under SPSS Categories [33]. All numerical attributes where treated as ordinal variables, such that nonlinear transformation was possible. However, due to the large flexibility of NL-PCA and the small data set, the resulting map has a lot of overlapping products. In fact, NL-PCA created clearly three clusters. Although the products in the clusters were also close together in the maps of the other three methods, they did not really form clusters. This example shows, in extreme, the ability of NL-PCA to cluster products having the same category value together.