Please see the Statistics Research Associates website for a formal CV.
I am interested in extracting information from data and in the statistical methods for doing this. On the theoretical side I have been interested in time-series methods and the use of large sample (asymptotic) methods in unusual situations. In these interests, I have been very much influenced by my first co-author, Professor David Vere-Jones, and by my graduate advisers Professors Jerzy Neyman and Lucien LeCam at Berkeley. And then by clients and colleagues in DSIR.
Some of my work has involved the manipulation of matrices and I have dreamed of a computer language that would allow the simple manipulation of matrices without the overheads and awkwardness of the standard statistical computing languages and the awkwardness in handling matrices in the regular computing languages. C++ seemed to allow the possibility of writing a nice matrix package, but it was much more complicated than I ever dreamed it would be.
Most of my real statistical analysis used to be carried out in the S-plus statistical language with help from Excel, Access and SQL server. However I do see a place for a statistical package embedded in C++ and am now carrying out some of my work in a preliminary version of this package.
Some of my recent work is involving getting data from quite large databases then using S-plus to get conclusions for clients. This involves both presenting the data graphically and using standard statistical analysis methods. I see this as an important area for new research: particularly, how do you do EDA graphics plus analyses such as fitting generalised linear models on these large databases?
Examples of my work - see also notes on road crash risk model.