Facet labelling with R

Question

I am trying to plots subsets of data using facet plot. It is 2 columns x4 row plots. First column with values between 120 and 150, and second column with values between 180 and 250, and the variables are assigned as Lesser or Greater in datafile. I am facing problem with adding of specific labels ('120 <= greek symbol(alpha) <= 150', '180 <= greek symbol(alpha) <= 250) to the column variables ("Lesser", "Greater"). I tried doing it as:

require(graphics)
library(ggplot2)
hp <- ggplot2(data) + ....
xlow <- paste("120 <", expression(alpha), " < 150") 
xhi <- paste("180 <", expression(alpha), " < 250")
mf_labeller <- function(var, value){
  value <- as.character(value)
  if (var=="regime") {
  value[value=="Lesser"] <- xlow
  value[value=="Greater"]  <- xhi
}
  return(value) 
}
(hp %+% data) + facet_grid(param~regime,  scales="free", labeller=mf_labeller)

I get "Error in value == "Greater" : comparison is not allowed for expressions". I tried coercing the data too. If I do the labelling as in any way:

data2 <- transform(data, regime = factor(regime, levels=c("Lesser", "Greater"), labels=c(expression(paste("120 <", alpha," < 150")), expression("180 < alpha < 250"))))

I get the whole expression verbatim, and not the desired plots. I may be missing something. Any help is really appreciated!! Thank you in advance..

The excerpts of my sample is:

regime,param,line,XX,Var,sner
Lesser,Rise,VII,AA,4.968624,0.1275248
Lesser,Rise,VII,BB,3.719405,0.08470305
Lesser,Rise,VII,CC,7.608773,0.177848
Lesser,Rise,VII,DD,9.874395,0.1367159
Lesser,Text,VII,AA,4.968624,0.1275248
Lesser,Text,VII,BB,3.719405,0.08470305
Lesser,Text,VII,CC,7.608773,0.177848
Lesser,Text,VII,DD,9.874395,0.1367159
Lesser,Chant,VII,AA,0.1771826,0.186758
Lesser,Chant,VII,BB,0.3611497,0.5484656
Lesser,Chant,VII,CC,0.7719002,0.8864444
Lesser,Chant,VIII,DD,1.829022,0.2639881
Greater,Rise,VII,AA,4.968624,0.1275248
Greater,Rise,VII,BB,3.719405,0.08470305
Greater,Rise,VII,CC,7.608773,0.177848
Greater,Rise,VII,DD,9.874395,0.1367159
Greater,Text,VII,AA,4.968624,0.1275248
Greater,Text,VII,BB,3.719405,0.08470305
Greater,Text,VII,CC,7.608773,0.177848
Greater,Text,VII,DD,9.874395,0.1367159
Greater,Chant,VII,AA,0.1771826,0.186758
Greater,Chant,VII,BB,0.3611497,0.5484656
Greater,Chant,VII,CC,0.7719002,0.8864444
Greater,Chant,VIII,DD,1.829022,0.2639881

and code I used: I wrote earlier expression in wrong way here in post, but i did use the correct one in code.

x <- read.table("sample.csv", header=T, sep=',')
require(graphics)
library(ggplot2)
ppi <- 300
png("figure.png", width=6*ppi, height=6*ppi, res=ppi)
hp <- ggplot(data=x,aes(x=XX, y=Var, colour=line, group = line)) + geom_errorbar(aes(ymin=Var-sner, ymax=Var+sner, colour=line), width=.3) + geom_line(aes(ymax=Var+sner), size=0.7) + geom_point(aes(ymax=Var+sner), shape=21, size=2,fill="white") + theme_bw() + theme(axis.text.x = element_text(angle=90,vjust=0.25), panel.grid.minor=element_blank(), panel.grid.major=element_blank(), panel.background =element_blank(), legend.position="none" ) + scale_colour_hue(l=40) 
xlow <- expression(paste("120 <",alpha," < 150"))
xhi <- expression(paste("180 <", alpha," < 250"))
.. earlier code block as alternative here...
data2 <- transform(x, regime = factor(regime, levels=c("Lesser", "Greater"), labels=c( bquote(120<.(alpha)~phantom()<150), bquote(180<.(alpha)~phantom()<250) )
(hp %+% data2) + facet_grid(param~regime,  scales="free", labeller=label_bquote)

I tried as earlier with mf_labeller too.. Thank you in advance

Brian Diggs · Answer 1 · 2012-11-20T21:30:18.943

The form you want for the plotmath expression is

expression(120 <= alpha~phantom() <= 150)
expression(180 <= alpha~phantom() <= 250)

which you can get that way or indirectly by

parse(text="120 <= alpha~phantom() <= 150")
parse(text="180 <= alpha~phantom() <= 250")

The easiest way is to set the levels of the factor to those strings and use the label_parsed function.

Mocking this up with the mtcars data, since you didn't provide a reproducible example:

dat <- mtcars   
dat$regime <- factor(dat$am)
levels(dat$regime) <- list("120 <= alpha~phantom() <= 150" = "0",
                           "180 <= alpha~phantom() <= 250" = "1")

ggplot(aes(x=wt, y=mpg), data=dat) +
  geom_point() +
  facet_grid(cyl ~ regime, labeller=label_parsed)

enter image description here

If you really want to use your own labeller, and don't want to change the factor levels of your underlying data, you can do that too. Start with mock data with "Lesser" and "Greater" levels:

dat <- mtcars   
dat$regime <- factor(dat$am, levels=c(0,1), labels=c("Lesser", "Greater"))

Then the labeller function and ggplot code would be:

mf_labeller <- function(var, value){
  value <- as.character(value)
  if (var=="regime") {
    value[value=="Lesser"] <- "120 <= alpha~phantom() <= 150"
    value[value=="Greater"]  <- "180 <= alpha~phantom() <= 250"
    value <- lapply(value, function(x) parse(text=x))
  }
  return(value) 
}

ggplot(aes(x=wt, y=mpg), data=dat) +
  geom_point() +
  facet_grid(cyl ~ regime, labeller=mf_labeller)

The graph looks the same.

score 0 · Answer 2 · edited May 23 '17 at 10:34

0

We do not have your data to test (nor do we have the code you are using), but prior experience with lattice plotting has taught me that either bquote or some combination of sapply and as.expression may be needed. Try :

..., labels=c( bquote(120<.(alpha)~phantom()<150), bquote(180<.(alpha)~phantom()<250) )

In the first section of code I would not have thought you should be using expression inside paste but rather constructing the labels first and then building an expression vector.

If you get desperate, you can build the plot first, and then force the labels to be something different. How can I use grid to edit a ggplot2 object to add math expressions to facet labels?

edited May 23 '17 at 10:34

Community

1
1

answered Nov 20 '12 at 20:35

IRTFM

258,963
21
364
487

Thank you for reply. But this also does not seem to work in my case. Though the code runs but no output plot is generated. – user1818565 Nov 20 '12 at 21:25
Sigh. Of course no output is generated. You provided no code or data. You will notice I hope that Brian Diggs used the phantom() strategy to get the plotmath to work. – IRTFM Nov 21 '12 at 00:06

MattBagg · Answer 3 · 2012-11-20T21:28:45.883

I don't have a real answer, but am putting this up in the hopes it is helpful. My solution throws an error with two less-than signs but works with one, which strikes me as buggy behavior. (Edit, see Brian Diggs comment below or answer presumably/hopefully above for why this isn't buggy.)

Sample data that works, note I am substituting a plus for one less-than in twolabs:

twolabs  <- c("120<alpha+150","150<alpha+250")
df<- data.frame(x=rnorm(mean=4,40),
   y=rnorm(mean=4,40),
   labs=sample(twolabs,40,replace=T))

This plot works beautifully.

ggplot(df,aes(x=x,y=y))+geom_point()+facet_grid(.~labs,labeller=label_parsed)

However, data with two less than signs throws an error:

twolabs  <- c("120<alpha<150","150<alpha<250")
df2<- data.frame(x=rnorm(mean=4,40),
   y=rnorm(mean=4,40),
   labs=sample(twolabs,40,replace=T))

> ggplot(df2,aes(x=x,y=y))+geom_point()+facet_grid(.~labs,labeller=label_parsed)
Error in parse(text = x) : <text>:1:10: unexpected '<'
1: 120<alpha<

By using label_parsed, I think we call ultimately call gridDevices::plotmath, for what it's worth. Can someone take this further?

`<` is a binary operator and each one needs its own left and right hand side (`a — Brian Diggs, Nov 20 '12 at 21:24

Facet labelling with R

3 Answers3