Cat_plot converts numeric predictor variable to categorical without warning

Using 'cat_plot' with a continuous predictor variable converts the predictor into a character variable. This means x axis labels can be very messy as they overlap and its difficult (or maybe impossible) to set breaks so these dont overlap. I had this issue but i eventually realised i should have been using 'interact_plot' instead of 'cat_plot'.

If 'cat_plot' tested for the variables being numeric before plotting and issued a warning that 'interact_plot' is better suited to plotting these variables it might save other users making this same mistake.

Thanks for a great package!

Start at the cat_plot entry point and trace how continuous predictor variables are handled before plotting. Compare the behavior with interact_plot; done means numeric predictors trigger a warning recommending interact_plot, with the existing plotting behavior otherwise understood and verified.

Written by the indexing model from the issue text.