These feature transformations are useful when you have categorical data, i.e., data where each observation is one of a discrete set of possible values.