ggplot renaming facet labels in facet_wrap

Question

I've hit a stumbling block in writing a ggplot function. I'm trying to change the facet labels in a ggplot facet_wrap plot.... but its proving trickier than I though it would be....

The data I am using can be accessed here

str(ggdata)
'data.frame':   72 obs. of  8 variables:
 $ Season : Factor w/ 3 levels "Autumn","Spring",..: 2 2 2 2 2 2 2 2 2 2 ...
 $ Site   : Factor w/ 27 levels "Afon Cadnant",..: 13 13 13 13 13 13 13 13 13 13 ...
 $ Isotope: Factor w/ 4 levels "14CAA","14CGlu",..: 1 1 1 1 1 1 2 2 2 2 ...
 $ Time   : int  0 2 5 24 48 72 0 2 5 24 ...
 $ n      : int  3 3 3 3 3 3 3 3 3 3 ...
 $ mean   : num  100 88.4 80.7 40.5 27.6 ...
 $ sd     : num  0 1.74 2.85 2.58 2.55 ...
 $ se     : num  0 1 1.65 1.49 1.47 ...

I have written the following function to create the ggplot which uses the Isotope factor levels to label the facets:

plot_func <- function(T) {site_plots <- ggplot(data = T) + geom_point(aes(Time, mean, colour = Season, shape = Season)) + 
  geom_line(aes(Time, mean, colour = Season, linetype = Season)) +
  geom_errorbar(aes(Time, mean, ymax = (mean + se), ymin = (mean - se)), width = 2) +
  labs(title = T$Site[1], y = "Percentage of isotope remaining in solution", x = "Time (h)") +
  scale_x_continuous(breaks=c(0, 24, 48, 72)) +
  scale_y_continuous(limits=c(0,115), breaks = c(0,25,50,75,100)) +  
  theme(axis.title.y = element_text(vjust = 5)) +
  theme(axis.title.x = element_text(vjust = -5)) + theme(plot.title =  element_text(vjust = -10)) +
  theme_bw() + facet_wrap(~Isotope, ncol =2) 
  print(site_plots)
  ggsave(plot = site_plots, filename = paste(T$Site[1], ".pdf"), 
     path = "C:/Users/afs61d/Dropbox/Academic/R/Practice datasets/Helens_data/Site_Isotope_Season_plots/", 
     width = 9, height = 7, dpi = 300)}

Resulting in this lovely graph:

Which is nice but I want to change the facet labels now... Having done some poking around google I thought I might be able to use the labeller function as an argument to pass to facet_wrap. After a frustrating hour I discovered that this only works with facet_grid!!!??? So, an alternative method was to change the Factor level names so give me the facet labels that I want::

 gdata$Isotope <- revalue(x = ggdata$Isotope, 
c("14CAA" = " 14C Amino Acids", "14CGlu" = "14C Glucose", 
  "14cGlu6P" = "14C Glucose-6-phosphate", "33P" = "33P Phosphate"))

This works, but the problem I have now is that I want the numbers in the labels to be super-scripted. Can anyone suggest the best way to achieve this? Thanks

Decay

I think "labeller" is available for facet_wrap in the development version of ggplot: docs.ggplot2.org/dev/facet_wrap.html – Domain

@BenBolker Thanks... that would be very useful - how can I install that? – Decay

@eipi10 I don't think labeller can be passed to facet_wrap... which is the crux of the problem – Decay

I've updated my answer to use facet_wrap, which works, but I'm using the development version of ggplot. – Trichology

related popular question but for facet_grid #3473480 – Elvia

Answer 1

Manage to sort it out! Had trouble installing the development version of ggplot but after installing curl and devtools and reinstalling scalesit worked. I tried @eipi10 answer but couldn't get that to work so I changed the factor label names in a different way:

ggdata$Isotope <- factor(ggdata$Isotope, labels = c("NULL^14*C~Amino~Acids", 
"NULL^14*C~Glucose", "NULL^14*C~Glucose-6-phosphate", "NULL^33*P~Phosphate"))

I then adjusted the ggplot function to pass labeller = label_parsed to the facet_wrap function:

plot_func <- function(T) {site_plots <- ggplot(data = T) + geom_point(aes(Time, mean, colour = Season, shape = Season)) + 
  geom_line(aes(Time, mean, colour = Season, linetype = Season)) +
  geom_errorbar(aes(Time, mean, ymax = (mean + se), ymin = (mean - se)), width = 2) +
  labs(title = T$Site[1], y = "Percentage of isotope remaining in solution", x = "Time (h)") +
  scale_x_continuous(breaks=c(0, 24, 48, 72)) +
  scale_y_continuous(limits=c(0,115), breaks = c(0,25,50,75,100)) +  
  theme(axis.title.y = element_text(vjust = 5)) +
  theme(axis.title.x = element_text(vjust = -5)) + theme(plot.title = element_text(vjust = -10)) +
  theme_bw() + facet_wrap(~Isotope, ncol =2, labeller = label_parsed) 
  print(site_plots)
  ggsave(plot = site_plots, filename = paste(T$Site[1], ".pdf"), 
     path = "C:/Users/afs61d/Dropbox/Academic/R/Practice datasets/Helens_data/Site_Isotope_Season_plots/", 
     width = 9, height = 7, dpi = 300)}

Passing the ggdata to the plot_func gives me the below graphs with the correct facet labels.

Decay

Answer 2

Set the facet labels to the appropriate expressions, then use the labeller function label_parsed to ensure that they are displayed properly. Here's an example, using the built-in iris data frame:

data(iris)
iris$Species = as.character(iris$Species)
iris$Species[iris$Species == "virginica"] = "NULL^14*C~Amino~Acids"

ggplot(iris, aes(Sepal.Width, Sepal.Length)) +
  geom_point() +
  facet_wrap(~ Species, labeller=label_parsed)

You need to add the NULL before ^14*C or you'll get an error due to having ^ as the initial character. * and ~ mark the boundaries of each part of the expression, depending on whether you don't or do want a space between each part.

As of this writing (Dec. 12, 2015), you need to use the development version of ggplot2 for this to work with facet_wrap. However, this feature will presumably soon be incorporated into a regular release of the package.

Trichology

neat trick with the "NULL" to enable the initial superscript. But the OP's question was about facet_wrap, not facet_grid. – Misguidance

@Trichology thanks but this wont work with facet_wrap... useful to see label_parsed in action though – Decay

I've updated to use facet_wrap, which also works. I'm using the development version of ggplot (ggplot2_1.0.1.9003). Maybe it doesn't work in the latest CRAN release? – Trichology

@Trichology great... how can I get hold of the development version? – Decay

If you don't have the devtools package, install it and load it. Then install_github("hadley/ggplot2"). – Trichology

@Trichology struggling with that getting the following message: Error in get(method, envir = home) : lazy-load database 'C:/Users/afs61d/Documents/R/win-library/3.2/ggplot2/R/ggplot2.rdb' is corrupt In addition: Warning messages: 1: In .registerS3method(fin[i, 1], fin[i, 2], fin[i, 3], fin[i, 4], : restarting interrupted promise evaluation 2: In get(method, envir = home) : restarting interrupted promise evaluation 3: In get(method, envir = home) : internal error -3 in R_decompress1 Error: package or namespace load failed for ‘ggplot2’ – Decay

This is happening when you try to install the dev version of ggplot2? – Trichology

@Trichology yes. Tried to download then it suggested I needed the curl package... so did that, which didn't work. So I updated from R3.2.2 to 3.2.3 then installed curl then installed the dev version of ggplot2. However, I hen get the above message when I run the ggplot function above – Decay

I'm not sure. Maybe you need to install cURL? – Trichology

Done that... not sure what has gone wrong. Will try and uninstall ggplot2 and start again – Decay

Answer 3

I guess you could also use something even simpler such as:

facet_wrap(factor(Species, levels = c("A", "B", "C"), labels = c("a", "b", "c")) ~.)

Rawboned