Very good article!
Can you please explain to me how you calculate 'probability_charged_off drift' and probability_charged_off average' shown in 'Identifying data drift causes' figure? The difference between those two lines is not clear to me.
Thanks!