Multivariate data sets are a result of numerous methods of analytical chemistry, for instance like high performance chromatography. In addition to procedures of nonsupervised learning (e.g. cluster analysis) multidimensional analysis of variance, discriminant analysis and classification methods have extraordinary importance for data evaluation. Essential ideas of these methods are described and their application to two data examples by using the BASIC programs VARDIS and CLASS is illustrated. The first example deals with a discrimination of fuel filling stations being provided by the same source. In another example the recognition rate of 10 european brandies is studied. Both data sets were obtained by gas chromatography.