so we have to predict the number captured each year i think.
id is structured like
CaptureSite_0_201901
CaptureSite_0 part is the location captured
2019 is the year
01 is what exactly? it cant be month because then it wouldnt go to 40 so what is it meant to represent
201901 is the 1st week in 2019 , 201902 is the 2nd and etc (1 year=52 weeks)
even then the data still doesnt match up because there only 44 weeks for each capture
For the years 1998 to 2018, you're given the data about the 52 weeks of each year , as for the submission file you're only asked to forecast the first 44 weeks of 2019.
thank you