I have this data
¦Year-Week¦Sales¦Forecast¦
¦2015-42 ¦ 7 ¦ ¦
¦2015-43 ¦ 8 ¦ ¦
¦2015-44 ¦ 2 ¦ ¦
¦2016-45 ¦4276 ¦ ¦
¦2016-46 ¦ 876 ¦ ¦
¦2016-47 ¦ 7 ¦ ¦
¦2016-48 ¦ 52 ¦ ¦
¦2016-49 ¦72876¦ ¦
¦2015-50 ¦ 20 ¦ ¦
¦2015-51 ¦ 35 ¦ ¦
¦2015-52 ¦ 15 ¦ ¦
¦2016-01 ¦ 24 ¦ ¦
¦2016-02 ¦ 8 ¦ ¦
¦2016-03 ¦ ¦ ¦
¦2016-04 ¦ ¦ ¦
¦2016-05 ¦ ¦ ¦
¦2016-06 ¦ ¦ ¦
¦2016-07 ¦ ¦ ¦
¦2016-08 ¦ ¦ ¦
I've tried this formula to workout the forecast figures for Year-Week 2016-03 onward.
=FORECAST(A6,$A$2:$A$5,$B$2:$B$5)
and got
#Value! --Error
After researching I have seen that excel will not recognise Year-Week column as it's non numeric, hence the error.
I then changed the Year-Week column to just display the Week and it has returned values however they look incorrect to me:
¦Year-Week¦Sales¦Forecast¦
¦42 ¦ 7 ¦ ¦
¦43 ¦ 8 ¦ ¦
¦44 ¦ 2 ¦ ¦
¦45 ¦4276 ¦ ¦
¦46 ¦ 876 ¦ ¦
¦47 ¦ 7 ¦ ¦
¦48 ¦ 52 ¦ ¦
¦49 ¦72876¦ ¦
¦50 ¦ 20 ¦ ¦
¦51 ¦ 35 ¦ ¦
¦52 ¦ 15 ¦ ¦
¦01 ¦ 24 ¦ ¦
¦02 ¦ 8 ¦ ¦
¦03 ¦ ¦39.75545¦
¦04 ¦ ¦39.15559¦
¦05 ¦ ¦39.15573¦
¦06 ¦ ¦39.15587¦
¦07 ¦ ¦39.15587¦
¦08 ¦ ¦39.15615¦
Is there away to workout forecast based on weeks? as I don't have year month day in the data set and only year weeks to work with, or is the above result something sensible to get back with the formula?