I think the concept is easier to understand by staying consistent with the `mpg` dataset, there's probably no need to introduce a new dataset for that remark only.
