search

jasp-stats / jasp-issues

This repository is solely meant for reporting of bugs, feature requests and other issues in JASP.
59 stars 29 forks source link

[Feature Request]: jamovi & Excel features for flexplot #2138

Open tomtomme opened 1 year ago

tomtomme commented 1 year ago

Still valid for 0.19 beta

Description

jamovi got some new UI features within flexplot that would be nice for JASP too. Same for Excel/Calc features.

Is your feature request related to a problem?

with the current UI flexplot in JASP cannot handle added variable plot, manual binning, very big data sets etc. See list in my last comment or this screenshot from flexplots current implementation in jamovi: grafik

dustinfife commented 11 months ago

flexplot can do repeated measures, but they have to be done as mixed models. Does that help?

dustinfife commented 11 months ago

If you have three measurements over time, mixed models works. Doing a "Difference plot" is highly prone to error: you have to a grouping variable indicating which belonged to Time 1 versus Time 2, and it has to be sorted, and you have to have identical numbers of measurement. An alternative is to create a difference score yourself and do a univariate plot of that.

Jittered-density plots are what I used to call beeswarm plots before I realized they had already been named.

As for the other two, I can consider adding these improvements. Thanks for the feedback!

tomtomme commented 11 months ago

If you have three measurements over time, mixed models works. Doing a "Difference plot" is highly prone to error: you have to a grouping variable indicating which belonged to Time 1 versus Time 2, and it has to be sorted, and you have to have identical numbers of measurement. An alternative is to create a difference score yourself and do a univariate plot of that.

I know about the problems of difference plots and the manual way how to get one. However just hitting a button is more convenient / faster than to calculate a difference score before.

Either way. Cool that you consider the improvements. Here a nice list to check them off:

More model fit options like in Excel:

grafik

Flexplot is already awesome and those would make it fabulous :D cheers, thomas

Reference: old duplicate about log and exp fits only https://github.com/jasp-stats/jasp-issues/issues/1779