I have a student project in which I would like to analyze application usage data specifically between three user groups. I have log data gathered including unique User IDs, actions and the identifier for the user group they belong to. Ex:

UserID   Action   Group  
  0001        A       1  
  0002        B       1  
  0003        A       2  
  0001        A       1  
  0004        B       1  

My problem now lies within analyzing and visualizing this data ideally in a manner that takes into account the amount of unique users within the user group. In this example if I were to do try and find a correlation between user group and likelihood of performing a specific action I would only get the output that a user in group 1 is twice as likely to perform action A than a user in group 2. Of course that is not correct, since group 2 has less users than group 1.

I would like to use jmp to do this analysis and the closest I could get was a contingency analysis which output a mosaic graph. Unfortunately the graph cannot take into account how many users are within each group.

I'm certain there must be some simple way of doing it, but I was unable to come up with a solution so far.

Any help would be much appreciated.


Edit: removed the timestamps as they weren't relevant.

  • $\begingroup$ How many possible actions & how many groups are there? Are you trying to incorporate the tome variable into the analysis? If so, how? $\endgroup$ – gung - Reinstate Monica Jul 13 '15 at 14:35
  • $\begingroup$ Hi, actually time doesn't factor in yet (I need that for a later analysis), sorry for the confusion. There are 7 actions and 3 groups. $\endgroup$ – Grinner Jul 14 '15 at 12:33

JMP has a "Freq" role option in each analysis launcher, and for contingency the frequency is accounted for in the mosaic plot. If that's not it and you want JMP-specific help, you can try the JMP Community forums.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.