This article presents some of the multivariate methods used in metabolomics, and addresses many of the data types and associated analyses of current instrumentation and applications seen from the point of view of data analysis.I cover most of the statistical pipeline – from pre-processing to the final results of statistical analysis (i.e. pre-processing of the data, regression, classification, clustering, validation and related subjects). Most emphasis is on descriptions of the methods, their advantages and weaknesses, and their usefulness in metabolomics. Of course, the selection of methods presented is not an exhaustive, but should shed some light on some of the more popular and relevant.