Non-Parametric Rank Correlations

[MKostura]

What inspired this wish list request?   At one time there was a nice add-in, Correlate Y by X Add-in that computed both the standard Pearsons correlation along with Non-Parametric fits.  It was very easy to engage and worked well.  However, that add-in has not worked on the newer versions of JMP.  It is very common to use the non-parametric rank and the Spearmans rho for many applications.  

What is the improvement you would like to see? 

Simply find a way to enable that add-in to work in JMP19 

Why is this idea important? 

Currently, the only way to access a non-parametric correlation to obtain a Spearmans rho value is the multivariate test platform.  That can work well if the number of variables is low.  But in many cases, the desire is to have a one by many correlation calculated.  And when the many are several hundred, the multivariate platform is simply inapproriate for that purpose.  

[scott_allen]

@MKostura - Thanks for submitting the wish. Do you have a reference to this Add-In or it's name? We can ask the author if it is something that could update and publish in the JMP Marketplace

[MKostura]

@scott_allen.  The add-in name was "correlate y by x"  and the Add-In ID was "Parris-CorrelateYbyX" The version was 0.07    I have a copy of the add-in but there is no information that provides the name of the author although I do recall it was a JMP employee.  Does not seem to be a way to attach to this message.

The market place is, IMHO, a disconnected mess.  Not obvious how to connect easily from the community (it should be front and center on the User Community home page menu).  I stumbled on it after several minutes on the site.  And the market place has scrubbed a lot of add-ins that I still commonly use. I understand some may be deprecated (like the correlate y by x) but others were useful still.