This module is concerned with the analysis of multivariate data, in which the response is a vector of random variables rather than a single random variable.

Part I of the module describes some basic concepts in Multivariate Analysis and then recaps and introduces some key ideas needed from linear algebra. Chapter 1 defines notation, introduces some datasets, and discusses exploratory data analysis. Chapter 2 provides a recap on some matrix algebra. Much of this will be familiar to you, but if not, we take the time to introduce the key mathematical concepts that will be relied upon during the module. Chapter 3 introduces matrix decompositions. We start with the spectral decomposition of square symmetric matrices (which you will have studied previously), and then introduce the singular value decomposition (SVD). The SVD is one of the most important concepts in this module, and is the key linear algebra technique behind many of the methods we will study.

A theme running through the module is that of dimension reduction. In Part II we consider three types of dimension reduction: Principal Components Analysis (in Chapter 4), whose purpose is to identify the main modes of variation in a multivariate dataset; Canonical Correlation Analysis (Chapter 5), whose purpose is to describe the association between two sets of variables; and Multidimensional Scaling (Chapter 6), in which the starting point is a set of pairwise distances, suitably defined, between the objects under study.

In Part III, we focus on methods of inference for multivariate data whose distribution is multivariate normal.

Finally, in Part IV, we focus on different methods of classification, i.e. allocating the observations in a sample to different subsets (or groups).

If you find any typos or mistakes, please email me at

Note: Although it appears like the notes are very long, this is because the R code, R output and plots, as well as the exercises and computer tasks, are embedded in the notes. The theory parts of the notes are less than 100 pages in total.