Hi,
Maybe I'm dreaming here, but I'm wondering if there is a way to have Excel identify patterns in multi-column data sets?
I'm trying to predict volumes (of people) based on historical data. I have a year of data with a number of people for each date.
I've manually identified day-of-the-week patterns (i.e. certain days are generally busier than others) and one other pattern based on week of the month (i.e. 4th wednesdays are busier than 2nd wednesdays). I did this essentially by hand; sorted then chopped data into day of the week chunks, and then which week (1st mondays, 2nd mondays, etc.) then ran basic distributions on those mini-chunks. So I have a nice table of:
(Min)(Mean-SD)(Median)(Mean+SD)(Max) for each Day category (1Mon,2Mon,3Mon.....3Fri,4Fri,5Fri)
Pattern-wise, nothing else jumps out at me, but I'm wondering if there are any built-in analytical tools that can cross-reference and help me spot anything else I'm missing.
I'm looking for something that can tell me e.g. "Mondays in January are slower than average mondays, but Tuesdays in January are busier than average Tuesdays" so to speak.
I'm already sorting it by day, by weekday, by week, by month and nothing else leaps out.
Thanks in advance!
Lenny
Bookmarks