In the project “Predicting House Sale Prices” the columns
['Alley', 'Fireplace Qu', 'Pool QC', 'Fence', 'Misc Feature'] contain more than 25% of
Dataquest drops those columns in their solution but in the documentation stays that
NaN in those columns means that the item is absent, i.e., the house doesn’t have an alley or a fireplace. So we remove some useful information for our model.