What inspired this wish list request?
For complex datasets with a lot of variables (columns) we often use the interactive graphbuilder to explore our data. However when it comes to the statistical evaluation I need to run the classical Fit X by Y tool and then run the mean comparison (t-test, HSD, etc) if I want to have a statistical analysis. In this case however I always need to define upfront my X and Y variables and need to go back to replot another X by Y variables if I need to.
What is the improvement you would like to see?
I would see a large productivity gain if the graph builder, or a similar statistical graph build would allow me to interactively let me build X by Y graphs and then to run statistical comparison analysis.
Why is this idea important?
As said, it would increase productivity, less time is needed and less complexity for everybody exploring larger complex datasets within JMP.