Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Find answers, ask questions, and share expertise about Alteryx Designer.

Linear Regression: Calculating partial R squared instead of running individual regressions



I have a workflow that makes a linear regression on 19 independent variables. :) 

What I want to do is rank them using partial R squared. To do so right now I have to run 19 linear regressions with the individual variables to get an R squared for each. 

When I used to do my statistical models in R I used the ppcor package to calculate a straightforward partial correlation (partial R squared). 


Has someone else had this problem? Does anyone have a smart way to do this or an R-based alteryx macro to calculate the partial R squared? 

Any advice is appreciated. 


Many thanks,



This capability isn't built directly into the Alteryx GUI. I would recommend you use the R Code Tool to take advantage of the ppcor package.

Do you have any idea how I can do this coming out of an Alteryx linear regression tool? :) 



Can you share a quick example of how you do this currently with R?