If given a dataset with thousands of variables, how would you determine which variables are most important?
