4 Replies Latest reply on Oct 4, 2016 10:42 AM by Dhanashree Arole

    Jitter Help

    Brian M

      Hi,

      Working on getting Jitter to work for a case where I have multiple items (companies) per city and the values overlay.  I've followed Alan's example but with multiple attemps can't get it to recognize there are multiple values in the cluster to Jitter.

       

      See attached example I'm following and my attempt (book2).

       

      Any assistance is appreciated.

      Thanks

        • 1. Re: Jitter Help
          Dhanashree Arole

          Brian,

           

          The jitter parameter is simply boolean true/false datatype. Cluster size is defined as WINDOW_MAX(INDEX())

           

          The jitter value seems to be always false. Can you explain setup to recreate the overlay?

           

          I will be glad to troubleshoot the calculations further.

           

          Thanks,

          Dhanashree

          • 2. Re: Jitter Help
            Brian M

            Dhanashree,

            That's what I'm struggling with - not sure why the value is always false but think the issue lies in the cluster size parameter.  It's not recognizing there is a cluster.  I've recreated the parameters and measures with direct copy from the example sheet and also by hand in case there were copy issues.  Neither worked.

             

            Re setup - in this case it's as simple as possible - name, lat, long.  I have city, state, country in the attached data set but to get the jitter to work it's not relevant.

             

            Thanks for looking

            • 3. Re: Jitter Help
              Brian M

              Getting a little closer.  Modified "IsJittered" and "Cluster Size" to compute using "Company Name" instead of "Table (across).  Now the points show the correct cluster size and jittered =  true. 

               

              Problem is the points are not spreading.  When checking the data the "new lat" and "new long" have the same values as the "pop lat" and "pop long".  Not sure why it's not calculating them.

               

              Changing the formula for "new lat" from

              IF [Jitter] THEN

              WINDOW_AVG( AVG([Pop Latitude]) ) + ([z_Spread Distance]*SIN([z_Spread Direction]))

              ELSE

              AVG([Pop Latitude])

              END

               

              to

               

              IF [Jitter] THEN

              WINDOW_AVG( AVG([Pop Latitude]) ) + (1)

              ELSE

              AVG([Pop Latitude])

              END

               

              shows it's entering the if block since the point moves.  [z_Spread Distance]*SIN([z_Spread Direction] seems to be 0 in each case.

               

              Suggestions?

              • 4. Re: Jitter Help
                Dhanashree Arole

                Is this something to do with the precision level of measures involved Brian?