If you were given a dataset that appears to have inconsistencies, how would you approach cleaning and analyzing it?