Computing new variables — recycling code

Discuss the jamovi platform, possible improvements, etc.

by jonathon » Wed Dec 02, 2020 7:27 am

If I use STATA or SPSS or R, I can send other researchers my raw data, plus syntax file and they can reconstruct every single step or change them.

this is also true of jamovi. you can return to any analysis you've created, any transform you've applied, or any computed variable you've made, and see exactly what was done. it's part of our design philosophy that nothing is lost, and everything can be returned to.

I think if it were possible to insert an R code chunk wherein one could write variables back to the dataset, and have that R code appear in the output pane for clarity, then it would go a long way to addressing this issue.

i think we're getting pretty close to being able to provide this feature. i'll try and have something to you in the next month.

User avatar
Posts: 1689
Joined: Fri Jan 27, 2017 10:04 am

by leehw » Wed Dec 02, 2020 5:04 pm

That would be very exciting! Thanks for working on this.
User avatar
Posts: 9
Joined: Fri Nov 22, 2019 8:48 am

by RaphaelHuber » Wed Dec 02, 2020 7:02 pm

This sounds awesome! Thank you for your efforts with Jamovi.

Yes you are right non-linear traceability really is already implemented in Jamovi, and I also must say, that the way you did it is truly awesome! Jamovi influences the way I start to dig into data because of its workflow in a very creative and playful manor. Your concept is so well thought out, that it really gets inconceivable how other statistics programs have not always done it the way you do it. It's so intuitive. You may not be apple people, but it feels to me kind of like the iPhone felt to me compared to phones before that.

So, in a way I can see, that spewing out linear code is sort of a step backwards for Jamovi. Yet the linear way is easier to reconstruct, so I'm really happy you work on this.
Posts: 2
Joined: Mon Nov 30, 2020 2:16 pm

by reason180 » Sun Feb 21, 2021 4:09 pm

Here's a two-step approach one might take:

Use R rather than jamovi to code and run all data transformations prior to running any actual analyses. Output the resulting data file, then load it into jamovi.

I'm thinking of doing this in a graduate class, in which I'll incorporate both R and jamovi. This would give students some knowledge of, and exposure to, using actual R code for computation, thus helping them learn R at a level more detailed than what they'd get from just looking at jamovi's R/jmv syntax.
Posts: 57
Joined: Mon Jul 24, 2017 4:56 pm


Return to General