Bivariate Data

This chapter deals with bivariate observations, i.e., those where datapoints consist of pairs of values (x, y) of numeric (discrete or, more commonly, continuous) variables. The variables can either represent raw features (e.g., height, weight, blood pressure) or be the result of a transformation from originally multivariate data via some dimensionality reduction procedure. The main tools for representing bivariate data are the scatter plots, which are also the focus of this chapter. Representing bivariate data by scatter plots plays a major role in exploratory data analysis, particularly for (1) qualitatively assessing the correlation between variables (or the lack of it) and (2) visualizing clusters. The chapter shows methods and applications.