Exploring Data in Three Dimensions


Contour Plots

In this section you create a contour plot. A contour plot assumes that the Z variable is functionally related to the X and Y variables. That is, the Z variable can be modeled as a response variable of X and Y.

A typical use of a contour plot is to visualize the response for a regression model of two continuous variables. If you model a response variable by using an analysis chosen from the AnalysisModel Fitting menu, you can add the predicted values of the model to the data table. Then you can create a contour plot of the predicted values as a function of the two regressor variables.

Contour plots are most useful when the X and Y variables are nearly uncorrelated. The contour plot fits a piecewise-linear surface to the data, which models Z as a response function of X and Y. The contours are level curves of the response function. By default, the minimum and maximum values of the Z variable are used to compute the contour levels.

The three variables in a contour plot must be interval variables.