If given a dataset with multiple variables, how would you determine which variables are most influential?