This is indeed a tricky question. As I cannot spoil it for other students with my answer I will just point to the Using numerical and categorical variables together notebook, where we mention
Sometimes object
data type could contain other types of information, such as dates that were not properly formatted (strings) and yet relate to a quantity of elapsed time.
As also implied by the explanation displayed after you have submitted your response, properly formatted dates require no extra transformation (I insist on the word “relate” as already containing the same numerical information) and can be directly used when training a model.
About the numerical_features
of question 5, this is cherry-picking of some yet not all the available numerical features in the dataset. The historical reason is that the selected subset of columns did not contain missing values (now none of the columns do).