Description

In many areas of science, particularly biology, the data obtained are naturally directions in two or three dimensions. Mathematically, these are values on the unit circle or sphere.

Statistically, they require different treatment to data lying on the real line. For example, it is not immediately obvious how one should compute the mean for such data and the central limit theorem does not directly apply in the usual way. A variety of special probability distributions have been developed to model directional data and inferential procedures have been adapted in various ways.

After developing an understanding of the basic concepts and statistical models for circular data, students may specialise into interesting areas of theory or application or into analysis of data using appropriate software such as R.

Prerequisites

Statistical Concepts II

It is strongly recommended that students doing this project should also take Statistical Methods III.

Resources

Wikipedia entry for Directional Statistics and links from it.

Fisher, Statistical Analysis of Circular Data , Cambridge University Press 1995 (ISBN 978-0521568906)

Fisher, Lewis and Embleton, Statistical Analysis of Spherical Data , Cambridge University Press 1993 (ISBN 978-0521456999)