topic Re: Calculate moving average for By-Groups in Discussions

Calculate moving average for By-Groups

JesperJohansen — Mon, 22 Jul 2013 10:51:37 GMT

I'm trying to calculate a moving average for a data-set containing several rounds of the continuous data. That is, I want a calculated average for every data-point, including the first and last. The average should be 7 steps in either direction, except for either end, where non-existing data beyond the 7 step window should simply ignored. Of course I don't want data form different By-Groups to be included in the same average. I was hoping to use the "Col Mean" function, but I can't figure out how to make the average "move". Any suggestions?

Re: Calculate moving average for By-Groups

keith — Mon, 22 Jul 2013 11:11:22 GMT

Do you have ETS licenced? If so, there is functionality within PROC EXPAND to calculate moving averages

Re: Calculate moving average for By-Groups

JesperJohansen — Mon, 22 Jul 2013 11:33:07 GMT

I do have access to EST, but I have never used it. I was hoping to do it directly in the formula editor.

Re: Calculate moving average for By-Groups

Jeff_Perkinson — Tue, 18 Oct 2016 20:17:46 GMT

Hi Jesper,

You can use the Summation function with an If() condition to do this.

Here's a picture of what it would look like in the formula editor.

And here's the text version:

Summation(

i = Row() - 7,

Row() + 7,

If( :DAY == :DAY

:DIAMETER,

)

) / Summation( i = Row() - 7, Row() + 7, If( :DAY == :DAY, 1, . ) )

Hope that helps.

-Jeff

Re: Calculate moving average for By-Groups

JesperJohansen — Tue, 23 Jul 2013 05:53:11 GMT

That was exactly what I needed. Great! Thank you.