I believe you're looking for the WINDOW_AVG function.

It will calculate the average of datapoints over the window you specify

So in your case, if you measure is "Results", and you've got the date you want as your dimension on the row or column shelf, you'd probably have something like:

WINDOW_AVG([Results],-9,0)

(whenever you use 0 as the start or endpoint in any of the WINDOW_### functions in Tableau, you will subtract 1 from the desired window size--hence -9 instead of -10.)

If it wasn't obvious, you'd create a new calculated field where you'd use this function.

Hi Emily,

Close.  I forgot to mention that with WINDOW_### functions, you have to specifiy an aggregation within them.

Usually, you want to use SUM because you are looking to get the moving average of the total.

Try this:

WINDOW_AVG(SUM([Weight per vat]),-9,0)

Does that get it?

Also, here's your workbook.

I will also mention a 2nd way to do the same thing.  It's perhaps simpler:

Observe the new "Sheet 2" in the attached workbook.

If you hold Ctrl and drag the "Sum(Weight per vat)" green Measure pill next to itself on the Rows shelf, Tableau will place a duplicate Measure for you. Then right click the duplicate pill and "Add Table Calculation."

You can choose the Calculation Type = "Moving Calculation"

for the Summarize values using: , choose "Average"

Then for Previous Values, you'd use 9 and next values 0 (check Include current value).

Since Moving Average is fairly common, Tableau tried to make it easy--this method will accomplish the same thing as creating the separate calculated field.

Hi Emily,

I re-read your original post, and I think I missed a detail. You wanted average results for each recipe.

That is not what I understood originally.

You said you want to average the last 10 runs, but am I correct that you do not care how many days ago the runs were, you always want to calculate the previous 10?

For example, the last 10 runs for HTST Recipe #10 were 6/27, 5/30, 5/9, 5/9, 4/18, 4/18, 3/21, 2/17, 1/18, and 12/28.

The problem with that is that there is no way in the underlying data to distinguish the 1st 5/9 run from the 2nd for Recipe #10. I don't know if it's possible within Tableau, but you could certainly just add one additional primary key column to the underlying data to allow additional differentiation. (so, just a "runID" that is an ascending integer, 1-2641 or more...)

Then, with an index on this "RunID" field, you will just want to average the results values for the 10 "greatest" ID's for a given recipe.  I have attached a workbook with the modified workbook and also with a new copy of your data with the field I'm talking about added.

Hope that's what you're after!