819077_en-US

liliconcom

Raking or Rim Weighting capability

Virtually every JMP platform allows for the possibility of using a Weighting variable as part of the modeling.  However, it would be nice if JMP also provided a capability for creating the weight variables themselves.  In survey data, sample populations may not "look" like the target populations.  Techniques like Raking (or other similar methods) could be used to create weights to rebalance the sample population to match its intended target population.  Both R and Python packages are out there that do this kind of thing.  Perhaps one could be converted into an addin or repurposed as an actual JMP platform. 

 

Automation and ScriptingConsumer and Market ResearchData Blending and CleanupRe: Raking or Rim Weighting capability - Status changed to: Acknowledged

Thank you for this suggestion @hengehold_da1, and for participating in the discussion about Raking and Rim Weighting last year. While we don't currently have plans to add this functionality directly in JMP, we like the suggestion of developing an add-in (perhaps using the new Python integration) and sharing via the JMP Marketplace.

 

FYI - Propensity Scoring and Propensity Score Matching are now available in JMP Pro 19 if you have a chance to try these out and provide feedback (via jmpbeta@jmp.com).

 

Thanks again! 

Re: Raking or Rim Weighting capability

Hi Mia, 

 

Thanks for your comments.  A couple of question, if you don't mind:

 

1. Are you suggesting that JMP likes the idea of creating an addin themselves and sharing it, or that someone else (who knows Python) should?
2. Where do I find the propensity scoring platform in 19EA?   I would like to take "observational" (vs. randomized) data where "treatment" may be affected by certain variables and create a weight variable ("weights") to account for any imbalances.  Is there a process for this in 19EA?

 

Thanks!