If you were given a large dataset with missing or inconsistent entries, how would you go about cleaning it?