API, as a tool for multivariate statistics

Abstract

This paper describes ways in which the combination of APL and graphics terminals can provide powerful means both of understanding and executing data analyses, and thereby lead to rapid discovery of structures and interconnections in multivariate data which numerical tabulation alone does not always reveal. The aims are to exploit the exceptional expository powers of APL (including APL2) to describe and differentiate between techniques currently used in multivariate analysis, and also to demonstrate how functions developed for exposition can be immediately applied in exploratory data analysis.