A scatter plot shows the relationship between variables.
The scatter plot identifies the relationship that best describes the data, whether a straight line, polynomial or some other function.
A scatter plot matrix shows the relationship between each predictor and the response, and the relationship between each pair of predictors. You can use the matrix to identify the relationship between variables, to identify where additional terms such as polynomials or interactions are needed, and to see if transformations are needed to make the predictors or response linear. The scatter plot matrix does not convey the joint relationship between each predictor and the response since it does not take into account the effect on the response of the other variables in the model. Effect leverage and residual plots fulfill this purpose after fitting the model.
