What inspired this wish list request?
I frequently use moving (rolling) fixed size window statistics that can be computed independently for an asset (byVar) for a number of X variables over a datetime ordered data. In my analysis, a predictive quality model has not only select important X's but also a datetime cycle where the effects are most significant from to distinguish normal from abnormal assets.
What is the improvement you would like to see?
A version of these exist in formula stack as Col moving average. I like the ability to add many other statistics using formulas (and not scripts).
I like the following moving statistics be available simply from formulas
these include (but not necessarily limited to) ..
moving specific quantiles (wish list from https://community.jmp.com/t5/JMP-Wish-List/Col-Moving-Quantile-function/idi-p/321573)
moving stdev
moving slopes
moving non parametric Kendall Tau statistic
one could imagine moving any statistic
Why is this idea important?
Without the formula functionality, I need to go back to a scripting to recall how to do this every single time I need it. I think one of the main strengths of JMP is being one of the easiest tool/platform to offer advanced statistics & modeling . Formulas is a great way to keep carrying the notion of "easiest". I can script but I prefer not to.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.