[Feature Request]: Automatic Transformation of Data with Johnson and/or Box Cox

[RolandWinkler72]

Description

I need to transform data automatically by using the Johnson or Box Cox transformation. JASP could also propose me the best transformation to use.

Purpose

To transform a set of data which are not normally distributed into normally distributed data

Use-case

I am Black Belt trainr for Safran Group. We are 20 trainers who have the same issue. We train approximately 80 Black Belts per year who face the same issue.

Is your feature request related to a problem?

No problem, but a more easy way to transform data than today (R-console or Python programming)

Is your feature request related to a JASP module?

No response

Describe the solution you would like

Having a set of data which are not normally distributed (after Shapiro Wilk test or normality check) I receive an advise from JASP which is the best transformation to use. I use this transformation as an automatic function in JASP and transform the data.

Describe alternatives that you have considered

R-console (quite complicated and not easy to demonstrate during my training, moreover I need to install dedicated packages and I must advise the students to do the same, as we are all members of a big compmany with strong IT restrictions, everybody should activate the IT admins to install)

Additional context

No response

[EJWagenmakers]

I agree we should have this (and I am a little surprised we do not already)

[JTPetter]

I believe in the "Compute new column" menu you will find these options:

Does this address what you were looking for?

Advice for transformation in certain places in the Quality Control module could of course be added where relevant.

[TarandeepKang]

Hi Both, There are some transformations available, but not the Johnson. Another way to automate some of this process of choosing among them, while also including a wider range of transforms, might be to use the bestNormalize package. https://cran.r-project.org/web/packages/bestNormalize/citation.html. What do you think? I don't advocate users applying this unthinkingly, but it could be a good heuristic/ start...?

[EJWagenmakers]

I'll look into this. But I do see YeoJohnson in the list

[TarandeepKang]

Sorry @EJWagenmakers I genuinely didn't see it there at the bottom.

[boutinb]

Which JASP version do you use? Because this function is there for quite a while. You have to scroll down the function list, as shown in the picture of @JTPetter

[TarandeepKang]

I have been keeping up with the latest versions, perhaps I just need some stronger glasses! Apologies again, Bruno.

[RolandWinkler72]

I believe in the "Compute new column" menu you will find these options:

Does this address what you were looking for?

Advice for transformation in certain places in the Quality Control module could of course be added where relevant.

Hello,
yes, I saw these functions in the "compute new column" but there are some points open for me (or I didn't get it solved)

• Is the yeojohnson exactly the johnson transformation?
• How do I get the parameters needed in the transformations (like "lambda"). I understood that the package JSU-Parameters could give them out but I want to avoid the programming in the R-console, better to use only the JASP functions easily.