TL;DR: This work aims to describe some of the more useful techniques in a way which it hopes will illustrate their connections with each other and with better-known statistical methods.
Abstract: Biological research in general, and medical research in particular, is characterised by the expensive nature of much of its experimental material. Once a subject is obtained, it is commonplace to take several different measurements. Multivariate analysis methods are thus particularly appropriate to biological data, and the fact that they are not very widely used is to some extent due to the way in which they are usually presented—as a disjointed collection of recipes in an elaborate mathematical notation. I aim to describe some of the more useful techniques in a way which I hope will illustrate their connections with each other and with better-known statistical methods.
TL;DR: In this article, an application of methods for analysis of biological data from preoperational-post-operational industrial surveys is presented, where repeated observations at the same sampling station form a time series in which the observations are not statistically independent, and the usual forms of statistical analysis do not apply.