Counting NaN values in pandas group by

Question

I have a df like this:

Country     product                 date_install                date_purchase           id
BR          yearly