In R, how can I make boxplot for multiple csv files and export as pdf files-CodePudding

I want to make boxplot for 80 csv files, the filenames look something like this: -NY_two.csv,CA_three.csv,FL_three.csv,....,NY_ten.csv.

Desirables include

(I) boxplot (export as pdf, 2 graphs per page)

See below for the 3 out of the 80 csv files


# All 80 files have the same column names - state, dept, year and revenue

#copy and paste to generate 3 out of 80 csv, 

# The datasets generated below represent 3 out of the 80 csv files

# Dataset 1
state <-c("NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY")

dept <- c("energy","energy","energy","energy","works",'works','works','works','fin','fin','fin','fin','parks','parks','parks','parks','trans','trans','trans','trans')
year <- c("two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two")
revenue <-c(1212.9,1253,1244.4,5123.5,1312,3134,515.8,2449.9,3221.6,3132.5,2235.09,2239.01,3235.01,5223.01,4235.6,2204.5,2315.5,6114,4512,3514.2)

NY_two <-data.frame(state,dept,year,revenue)


# Dataset 2
state <- rep("FL",20)

dept <- c("energy","energy","energy","energy","works",'works','works','works','fin','fin','fin','fin','parks','parks','parks','parks','trans','trans','trans','trans')
year <- rep("three",20)
revenue <-c(112.9,123,124,523.5,112,334,55,449,221.6,332,235,239,235,223,235.6,204,315.5,614,512,514.2)

FL_three <- data.frame(state,dept,year,revenue)

# Dataset 3
state <- rep("CA",20)

dept <- c("energy","energy","energy","energy","works",'works','works','works','fin','fin','fin','fin','parks','parks','parks','parks','trans','trans','trans','trans')
year <- rep("three",20)
revenue <-c(1102.9,1023,1024,5203.5,1012,3034,505,4049,2021.6,3032,2035,2039,2035,2023,2035.6,2004,3015.5,6014,5012,5014.2)

CA_three <- data.frame(state,dept,year,revenue)

# exporting the the above datasets as csv files ( imagine them as 3 out of the 80 files)
# set the path in the write.csv(/path/.csv) to collect the datasets

write.csv(NY_two,"C:\\Path to export the DataFrame\\NY_two.csv", row.names = FALSE)
write.csv(FL_three,"C:\\Path to export the DataFrame\\FL_three.csv", row.names = FALSE)
write.csv(CA_three,"C:\\Path to export the DataFrame\\CA_three.csv", row.names = FALSE)

My attempt

# Desirables include
#(I) plot the boxplot & export as pdf file (2 graphs per page)

######################################################################################

library(ggplot2)

# import all csv files in the folder
files <- list.files("C:\\path to the files\\", pattern="*.csv", full.names = T)
files

# set the pdf file path, I want two plots per page
pdf(file = "/Users/Desktop/boxplot_anova.pdf")

#specify to save plots in 2x2 grid
par(mfrow = c(2,2))

out <- lapply(1:length(files), function(idx) {
  # read the file
  this_data <- read.csv(files[idx], header = TRUE) # choose TRUE/FALSE accordingly
  # boxplot using ggplot
   p <-ggplot(this_data, aes(x = dept, y = revenue, fill = dept))   
       stat_boxplot(geom = "errorbar", width = 0.15)   geom_boxplot(alpha = 0.8,    # Fill transparency
                                            colour = "#474747",   # Border color
                                            outlier.colour = 1)  theme(panel.background = element_blank())  ggtitle("Title using each file name ")

  p
dev.off() 
})

out

Kindly share your code, thanx in advance

CodePudding user response：

There are a few separate issues that might cause problems in your code:

generating plots in a function might not be properly exported (use plot(p) or print(p) instead of p).
You have to open the pdf device before your loop and close it after, not within the loop. E.g. this would work in principle:

pdf(file = "boxplot_anova.pdf")
#specify to save plots in 2x2 grid
par(mfrow = c(2,2))
out <- lapply(1:length(files), function(idx) {
  # read the file
  this_data <- read.csv(files[idx], header = TRUE) # choose TRUE/FALSE accordingly
  # boxplot using ggplot
   p <-ggplot(this_data, aes(x = dept, y = revenue, fill = dept))   
       stat_boxplot(geom = "errorbar", width = 0.15)   geom_boxplot(alpha = 0.8,    # Fill transparency
                                            colour = "#474747",   # Border color
                                            outlier.colour = 1)  theme(panel.background = element_blank())  ggtitle("Title using each file name ")
   plot(p)
})
out
dev.off()

The code above will not plot (up to 4, which you would expect from mfrow(2,2)) plots on the same page, since ggplot2 does not use base graphics. Use e.g. the plot_grid function from the cowplot package to achieve this. To generate multiple pages, split the plot list in matching number of elements, e.g. for 4 plots per page:

res <- lapply(files, function(x){
    this_data <- read.csv(x, header = TRUE) # choose TRUE/FALSE accordingly
  # boxplot using ggplot
   ggplot(this_data, aes(x = dept, y = revenue, fill = dept))   
       stat_boxplot(geom = "errorbar", width = 0.15)   
       geom_boxplot(alpha = 0.8,    # Fill transparency
           colour = "#474747",   # Border color
           outlier.colour = 1)  
       theme(panel.background = element_blank())   
       ggtitle(gsub("(.*/)(.*)(.csv)", "\\2", x))
})

# set the pdf file path, I want two plots per page
pdf(file = "boxplot_anova.pdf")
lapply(split(res, ceiling(seq_along(res)/4)), 
    function(x) plot_grid(plotlist=x, ncol=2, nrow=2))
dev.off()