Thankyou so much @ckrapu for your time !
Yeah, of course, here is my problematic script → divergences_binomial.py (2.0 KB)
and my little dataset → cuenta_normal_añoentero1.csv (4.2 KB)
mm I think the idea of dividing the column of days by weeks of the year is interesting, however the weight that will really show me whether a movement is more (or less) likely to be a supermarket movement is the quantity CANTIDAD, because the days are equally likely for all its range [1,30] to have a supermarket movement (so that indicates me that in the linear formula CANTIDAD will have strong weight compare to DIA).
I hope we can find a way to resolve these divergences.
Again, thank you very much for your help 