Produce multiple ggplot figures within one ggplot()

Question

I would like to use the same ggplot code to produce 8 different figures conditional upon figures in my dataframe. Usually I would use facet_grid, but in this case, I would like to end up with a pdf of each individual figure. For example, I would like one pdf for each row here:

df <- read.table(text = "
xvalue     yvalue    location    planting    crop
  1          5          A          early      corn
  2          3          A          late       corn
  6          2          A          early      soy
  7          4          A          late       soy
  4          7          S          early      corn
  2          6          S          late       corn
  3          2          S          early      soy
  5          1          S          late       soy
", sep = "", header = TRUE)

Basic ggplot:

library(ggplot2)

ggplot()+
  geom_point(aes(x=xvalue, y=yvalue), data=df)

but instead of facet_grid to get the location x planting x crop combos, I want one separate pdf of each.

Something [like this](http://stackoverflow.com/a/19147917/2461552)? — aosmith, May 18 '15 at 20:36

score 1 · Answer 1 · answered May 21 '15 at 09:54

Here's another approach:

library(plyr)
library(ggplot2)

df <- read.table(text = "
xvalue     yvalue    location    planting    crop
  1          5          A          early      corn
  2          3          A          late       corn
  6          2          A          early      soy
  7          4          A          late       soy
  4          7          S          early      corn
  2          6          S          late       corn
  3          2          S          early      soy
  5          1          S          late       soy
", sep = "", header = TRUE)

fplot <- function(d)
  {
     require(ggplot2)
     p <- ggplot(d, aes(x = xvalue, y = yvalue)) +
             geom_point(size = 4) +
             xlim(0, 10) + ylim(0, 10)
     file <- with(d, paste0(paste(planting, crop, location, sep = "_"), ".pdf"))
     ggsave(file, p)
  }

d_ply(df, ~ rownames(df), fplot)

The file names look like early_corn_A.pdf and are saved in your current working directory. I set fixed x/y limits for visual convenience. The function takes a one-row data frame as input and outputs a plot in pdf form. The plyr::d_ply() function processes each row of the data frame and produces a separate plot per row.

black_sheep07 · Accepted Answer · 2015-12-11T13:35:05.603

First I made your MWE into a data.table because it's faster

library(data.table)
library(ggplot2)
library(gridExtra)

df <- data.table(read.table(text = "
            xvalue     yvalue    location    planting    crop
            1          5          A          early      corn
            2          3          A          late       corn
            6          2          A          early      soy
            7          4          A          late       soy
            4          7          S          early      corn
            2          6          S          late       corn
            3          2          S          early      soy
            5          1          S          late       soy
            ", sep = "", header = TRUE))

I pasted together your planting and corn information to create a separate column:

df[ , plantingCrop := paste(df$planting, df$crop, sep = "-") ]

I created an character vector that has all combinations of planting and crop. You'll see why in a second:

plantingCrop1 <- unique(df$plantingCrop)

I use gridExtra to create all of the plots in separate .pdf pages. I basically created a loop that plots as many graphs as there are characters in the plantingCrop1 object I made above. In each loop, dat is the subsetted group that you want to plot, by the unique plantingCrop group from when we pasted the planting and crop columns together. It repeats this until everything is done.

pdf("plantingCrop.pdf", onefile = TRUE)
for(i in 1:length(plantingCrop1)){
        dat <- subset(df, plantingCrop==plantingCrop1[i])
        cropPlot <- ggplot(dat, aes(xvalue,yvalue)) + 
                    geom_boxplot(aes(color = location)) + 
                    theme_bw() + 
                    ggtitle(bquote(atop(.("Boxplot of Stuff"), 
                          atop(italic(.(plantingCrop1[i])), "")))) +
                    labs(x = "xvalue", y = "yvalue") +
                    theme(legend.position = "top", legend.title=element_blank())
        grid.arrange(cropPlot)
        }
dev.off()

I also included the right way to name the files using the plantingCrop name as a subtitle. That's in the ggtitle call.

I'd encourage you to change geom_boxplot(aes(color = location)) to geom_boxplot(aes(fill = location) when you have more data as it shows up on a graph better, but I left it that way for now so you can see the different groups.

score 0 · Answer 3 · answered Oct 12 '15 at 17:40

0

I came across a similar problem.

I simply used

gridExtra()

Available here

Then, I used another package named :

cowplot()

Available here

Then I followed the steps here.

And it worked like a charm. :D

Hope it helps.

answered Oct 12 '15 at 17:40

Pragyaditya Das

1,648
6
25
44

Produce multiple ggplot figures within one ggplot()

3 Answers3