jamovi

Dear all,

thank you for creating jamovi!!! I am just starting to use it and I have heard that it supports data in long format. However, I used a perfectly fine long format dataset and it doesn't seem to give out the right descriptive statistics. So my simple question is? Does jamovi support long format? If so, in which analyses? can one restructure data from wide to long and vice versa? If not, when is this coming. Data handling techniques such as reshaping are probably what we need the most. So this question is really pressing for us. Thank you very much. Dear wishes, Till

ps I couldn't find a similar topic with the search function so I created a new topic

hi,

sorry, no data shaping at present. the only analyses that support long format are the ones in GAMLj sorry. you can use the filters to remove duplicate rows, but i appreciate that's a bit of a hack.

probably won't be added very soon either, sorry. it's a big job.

jonathon

Hi Jonathon,

thank you so much for your reply, it's such a great spirit that you actually answer to the community, I really like the idea of Jamovi and hope that it will be further promoted!!! I totally understand it is a big job, however, it would be so worthwhile to think about implementing it sooner rather than later imho, as features such as those from the tidyverse, for data wrangling would be so meaningful for experimental psychologists (and for others too I assume). We spend most of our data time with preparing datasets to make them ready for analyses. Wide to long reshaping, aggregating, choosing rows and columns, all this stuff would be amazing to have it in Jamovi for students to access more easily than R. It would be so great if students could take over the job of dataset preparation more easily, and if they could do so in Jamovi as Jamovi is our choice that we teach them when they want to avoid R.
Thank you for your consideration
All the best
Till

Hi there!

First of all, thanks for all your work, jamovi really is an awesome tool!

Please don't think that I want to inappropriately promote my own tool, however, I created a tool within my ShinyApp that allows the user to upload their data, create subsets, trim their data (fixed values or condition means and SDs) and then transform these data from long to wide format, calculating condition means/medians and aggregated raw error rates and optionally arcsine transformed error rates (due to my background in experimental psychology and behavioral experiments).
The tools could be found here: https://chbuesel.shinyapps.io/experimenterstools/ (under the tab "Data Management") and the source code is on Gitlab (https://gitlab.com/chbuesel/experimenterstools).

I would love to contribute these tools to jamovi, however, as you might see from my source code, I am not the most competent programmer and more concerned that something works. I guess that a similar procedure in jamovi would require the option that (a) users can download the transformed data and open them in jamovi, or (b) that the long-format data is replaced by the transformed data directly in jamovi. However, I am afraid that I don't know how to accomplish that but i would be very willing to help implementing such a tool. If anyone would be up for that, please let me know.

Thanks again!

Best,
Christian

hey christian,

so if you could make your tool produce an URL that could be copy/pasted into jamovi (so jamovi can download and open it), i think this could work.

(in jamovi, when you go file -> open, there's a search box you can type into. you can paste URLs into that, and jamovi will open them).

would that be possible?

cheers

jonathon

Hi there,

sorry for taking so long. I condensed the tools relevant to this topic down into a small shiny app (https://chbuesel.shinyapps.io/datapreparation/). However, this link cannot be opened in jamovi, leading me to slowly suspect that I might have misinterpreted your earlier post. However, I hope that this tool is still useful for some.

Best,
Christian

hey christian,

so the way it could work is someone goes to your shiny app, reshapes their data, and then clicks "save" or "export" or something. the data is then saved somewhere (on a server? to an S3 bucket? somewhere), and then the user is provided with a link.

then they can copy/paste this link into jamovi, jamovi goes and retrieves the file, and voila, they're in business.

i appreciate this goes a little beyond what shiny apps usually do. i'm happy to give you some more pointers if this is a road you want to go down.

cheers

jonathon

Hi Jonathon,

Thank you for your quick reply and your already very useful pointers. May I just ask why a S3 bucket or something similar would be more useful than the currently used way (namely reshaping the data, download them locally as a csv and then open them in jamovi)?

Sorry if that's a silly question, I'm just relatively new in developing applications like that.

Best,
Christian

it's all about reducing friction. people like it when you reduce the number of steps involved. once we launch the online version of jamovi, it becomes even more compelling -- clicking a link in your app can take them straight to jamovi and automatically open that data set for them.

or clicking a link in jamovi pops up your app

but yup, downloading, opening, that will work too!

jonathon

Thanks a lot for your response! I understand and I'll try to implement it like you suggested. I hope it'll work reasonably soon.

Best and thanks again,
Christian