Subscribe Bookmark RSS Feed

Calculate moving average for By-Groups

JesperJohansen

Community Trekker

Joined:

Apr 15, 2013

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?

BR
Jesper
1 ACCEPTED SOLUTION

Accepted Solutions
Solution

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.

3927_JMPScreenSnapz002.png

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

-Jeff
4 REPLIES
keith

Community Member

Joined:

Feb 24, 2012

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

JesperJohansen

Community Trekker

Joined:

Apr 15, 2013

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

BR
Jesper
Solution

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.

3927_JMPScreenSnapz002.png

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

-Jeff
JesperJohansen

Community Trekker

Joined:

Apr 15, 2013

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

BR
Jesper