CODEWITHSUNDEEP
rcheckboxggplot2
crysis405
crysis405

Reputation: 1131

r checkboxes {manipulate} to toggle data series on and off

I'm trying to draw a graph which has three lines using ggplot. But then I want to be able to toggle the lines on or off so that when I', looking at an overcrowded plot I can just hide some of the data series. I can't seem to figure out how to do this (or even if its possible) with rstudio's manipulate package which includes checkboxes. Can someone please show me how to do this with this package or any other? Thank you

Code:

manipulate((ggplot(MeanFrameMelt, aes(x=variable, y=value, 
           color=id))+ geom_point()), 
           id = checkbox(FALSE, "File1"))

MeanFrameMelt (data):

id  variable    value
1   file1   V1  0.04114207
2   file2   V1  0.31830645
3   file3   V1  0.05797068
4   file1   V2  0.04138554
5   file2   V2  0.31510753
6   file3   V2  0.05830449
7   file1   V3  0.04157882
8   file2   V3  0.31220430
9   file3   V3  0.05865419
10  file1   V4  0.04177334
11  file2   V4  0.31117608
12  file3   V4  0.05900918

Upvotes: 0

Views: 1138

Answers (3)

Greg Snow
Greg Snow

Reputation: 49650

Here is one approach that uses the tkexamp function from the TeachingDemos package rather than manipulate:

library(TeachingDemos)
tklist <- rep( list(list('checkbox',init="T")), 3 )
names(tklist) <- levels( MeanFrameMelt$id )

tkfun <- function(...) {
    w <- c(...)
    w2 <- names(w)[w]
    df <- MeanFrameMelt[ MeanFrameMelt$id %in% w2, ]
    print(ggplot(df, aes(x=variable, y=value,
                  color=id)) + geom_point() )
}

tkexamp( tkfun, tklist )

Or it looks a little nicer if the last line is:

tkexamp( tkfun, list(id=tklist), plotloc='left' )

Someone else with more familiarity with ggplot2 will need to chime in on how to modify this so that the colors stay the same and optionally the y limits stay the same.

Edit

Here is a version of the function to keep the colors and the y limits the same even when not displaying some of the values:

tkfun <- function(...) {
  w <- c(...)
  w2 <- names(w)[w]
  df <- MeanFrameMelt[ MeanFrameMelt$id %in% w2, ]
  print(ggplot(df, aes(x=variable, y=value, colour=id)) + geom_point() +
    scale_colour_discrete(drop=FALSE) + 
    ylim(range(MeanFrameMelt$value))
  )
}

Upvotes: 1

jverzani
jverzani

Reputation: 5700

This is not really a manipulate question, but rather a question of you have 3 Boolean values to use when creating your plot expression. How to do it? The checkbox call creates these. The following doesn't exactly answer your specific question, but shows how to make this less confusing -- just add the lines one by one at the expense of being elegant. This could be done with RStudio, but I'm using a version of manipulate that works without RStudio:

require(gWidgets2) ## from github
options(guiToolkit="Qt") ## other choices too
source(system.file("examples", "manipulate.R", package="gWidgets2"))

manipulate({
  plot(mpg ~ wt, mtcars)
  if(do_lm)
    abline(lm(mpg ~ wt, mtcars))
  if(do_loess)
    with(mtcars, lines(lowess(wt, mpg)))
  ## ...
},
           do_lm=checkbox("Add regression line"),
           do_loess = checkbox("Add lowess fit")
           )

Upvotes: 2

Carl Witthoft
Carl Witthoft

Reputation: 21532

Here's one thing I figured out, based on Rstudio's example bar plot:

manipulate(
   barplot(as.matrix(longley[,factor]), 
           beside = TRUE, main = factor),
   factor = picker(1,2,2:3))

If you select the third option, 2:3 , it will plot both sets of data. Presumably the same sort of trickery can be used in other plot types. I'd be interested in seeing what others can come up with.

Update: the following works, albeit a bit clunky:

manipulate(matplot(foo[,1],foo[,c(which(c(fp2,fp3)==1))],t='l'), 
           fp2 = checkbox(TRUE,'col2') ,
           fp3 = checkbox(TRUE,'col3'))

You'll need an fp4 line if you want 3 checkboxes per your comment.

Note -- I have a question posted in Rstudio's forum because I got error messages when trying to assign checkbox outputs to a vector, i.e. fpick[4]=checkbox(TRUE,'col4')

Upvotes: 0

Related Questions