Calculator Options

Discuss the jamovi platform, possible improvements, etc.
Post Reply
User avatar
BobMuenchen
Posts: 16
Joined: Thu Jan 18, 2018 6:24 pm
Location: Knoxville
Contact:

Calculator Options

Post by BobMuenchen »

The JMP formula editor is a user interface masterpiece. It might provide some useful ideas as you expand jamovi's capabilities. As you can see from this 3-minute video, the basics are not all that different from jamovi now. It’s just additional eye-candy: https://youtu.be/HZCZ9b9WYUI.

Where it really shines is with conditional formulas. SPSS’ GUI keeps conditionals very simple by limiting the logic to select one subset at a time. However, that prevents it from being able to create a single formula for the entire transformation, so it wouldn't work for jamovi. JMP’s calculator handles conditionals much better by either letting you specify a unique set of logic for each condition, or by using a match function and pre-filling the conditions. The latter meets quite a lot of conditional needs, such as dietary calculations differing by gender. It would go find how gender is coded & fill those values in, then have a blank for the formula for each. I couldn’t find a video of that, but here’s a description of how it works.

https://www.jmp.com/support/help/13-2/U ... tml#165039

JMP is particularly fast at recoding values since it provides the values that it finds in the variable. Here’s a 3-minute demo showing how it does that:

https://www.youtube.com/watch?v=oADpjz8uSPE

Here's their master documentation: : https://www.jmp.com/support/help/13-2/F ... itor.shtml#
User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: Calculator Options

Post by jonathon »

you're right about that JMP formula editor! it's pretty nice! (but as you say, doesn't add *that* much).
Where it really shines is with conditional formulas. SPSS’ GUI keeps conditionals very simple by limiting the logic to select one subset at a time. However, that prevents it from being able to create a single formula for the entire transformation, so it wouldn't work for jamovi
ah yup. we are going to have a pretty pimped out recoding facility. don't worry about that! :)

but thanks for attending us to these things. very helpful.

jonathon
User avatar
BobMuenchen
Posts: 16
Joined: Thu Jan 18, 2018 6:24 pm
Location: Knoxville
Contact:

Re: Calculator Options

Post by BobMuenchen »

Yeah, it's really just their conditional setup that really shines. When I have to show someone a set of nested ifelse statements in R, their head explodes (unless they were already a decent programmer). Here's the example I usually start with since all instructors can relate. With JMP, it's a single simple dialog to fill in.

# The rep function repeats values.
grade <- rep("F", length(score))
grade

# Now they earn their way to a better grade.
# Each FALSE condition goes to a new nested ifelse.
# The last ifelse gets "grade" if all others are false.
grade <-
ifelse(score >= 90, "A", # False goes to new line...
ifelse(score >= 80, "B", # These are nested ifelse's.
ifelse(score >= 70, "C",
ifelse(score >= 60, "D", grade))))
User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: Calculator Options

Post by jonathon »

although, there is that dplyr function ... but yeah, i know what you mean.
Post Reply