loop Freq plot with GGPLOT aes_string

问题:

I have reviewed the previous question on stack overflow that relate to my ggplot question, but I was unable to find something that clearly helps.

Question: How can I modify the code below to generate separate frequency plots (histograms) for each column (variable) in the data frame using a loop. i.e. ID x each variable?

Data:

example.xlsx

ID  a1.sum  b3.sum  c6.sum  d9.sum
April Showers   10  5   15  0
Anita Job   2   3   1   14
Candy Cain  4   7   14  17
Crystal Ball    6   8   16  12
Dot Matricks    15  9       1
Kay Largo   4   10  5   13

Code:

#set work DIR
setwd("C:/A")

library(rJava)
options(java.parameters = "-Xmx2048m")  ## memory set to 2 GB

library(xlsx)

#read in .xlsx file and apply encoding UTF-8 (French accents)
DAT <- read.xlsx("example.xlsx", 1, encoding="UTF-8")


#plot data
library(ggplot2)   

p <- ggplot(subset(DAT, a1.sum>1), aes(ID, a1.sum, y=a1.sum))    
p <- p + geom_bar(stat="identity", fill="blue", color="green")
p <- p + theme(plot.background = element_rect(fill = "white"),
           panel.background = element_rect(fill = "white"),        
           panel.grid.major = element_line(colour = "white",size=0.25),
           panel.grid.minor = element_blank())
p <- p + theme(axis.text.x=element_text(size=10,angle=90, hjust=1, face="plain", family="serif"))  
p <- p + theme(axis.text.y=element_text(size=10, hjust=1, face="plain", family="serif"))
p <- p + theme(axis.line.x = element_line(color="black", size = 0.50),
           axis.line.y = element_line(color="black", size = 0.5))
p
ggsave(filename="a1.png", plot=p)

Output:

Plot of a1.sum Example of plot output

Trying to create a loop to generate the same plot for variables b3, c6, and d9.

I have tried several different approaches using aes_string. The following is how I am trying to setup the loop:

#get variable names that end in .sum
n <- names(DAT[grep("*.sum",names(DAT))])

#loop through variable names
for (i in 1:length(n)){
  in_dat <- c(n[i])

   ...ggplot...

print(p[i]);

}

回答1:

Original Answer - Using Facet Wrap

This sounds like an opportunity to use facet_wrap in ggplot2. You can first gather your data using tidyr in order to go from a wide format to a narrow format. Also, I used read.table based on your data and one row was missing a value so I filled that with 0.

DAT <- read.table(text = "ID  a1.sum  b3.sum  c6.sum  d9.sum
April_Showers   10  5   15  0
Anita_Job   2   3   1   14
Candy_Cain  4   7   14  17
Crystal_Ball    6   8   16  12
Dot_Matricks    15  9   0    1
Kay_Largo   4   10  5   13", 
                 header = TRUE, stringsAsFactors = FALSE)

    library(tidyr)
#gather data with
df2 <- gather(DAT, key, value, -ID)

This gives us:

> df2
              ID    key value
1  April_Showers a1.sum    10
2      Anita_Job a1.sum     2
3     Candy_Cain a1.sum     4
4   Crystal_Ball a1.sum     6
5   Dot_Matricks a1.sum    15
6      Kay_Largo a1.sum     4
7  April_Showers b3.sum     5
8      Anita_Job b3.sum     3
9     Candy_Cain b3.sum     7
10  Crystal_Ball b3.sum     8
11  Dot_Matricks b3.sum     9
12     Kay_Largo b3.sum    10
13 April_Showers c6.sum    15
14     Anita_Job c6.sum     1
15    Candy_Cain c6.sum    14
16  Crystal_Ball c6.sum    16
17  Dot_Matricks c6.sum     0
18     Kay_Largo c6.sum     5
19 April_Showers d9.sum     0
20     Anita_Job d9.sum    14
21    Candy_Cain d9.sum    17
22  Crystal_Ball d9.sum    12
23  Dot_Matricks d9.sum     1
24     Kay_Largo d9.sum    13

Then we make the same plot as before but it will be split by the key column. I have noted where I made changed below.

library(ggplot2)

p <- ggplot(df2, aes(x = ID, y=value))    ###Change made here
p <- p + geom_bar(stat="identity", fill="blue", color="green")
p <- p + theme(plot.background = element_rect(fill = "white"),
               panel.background = element_rect(fill = "white"),        
               panel.grid.major = element_line(colour = "white",size=0.25),
               panel.grid.minor = element_blank())
p <- p + theme(axis.text.x=element_text(size=10,angle=90, hjust=1, face="plain", family="serif"))  
p <- p + theme(axis.text.y=element_text(size=10, hjust=1, face="plain", family="serif"))
p <- p + theme(axis.line.x = element_line(color="black", size = 0.50),
               axis.line.y = element_line(color="black", size = 0.5)) +
  facet_wrap(~key) #facet added here

Updated Answer - Creating Separate ggplot Objects

In order to create a list of ggplot items, I borrowed heavily from this question. You create a function which you can then pass to lapply to make the plots.

First, make the function:

make_plots = function(data, column){
  ggplot(data, aes_string(x = "ID", y=column)) +
  geom_bar(stat="identity", fill="blue", color="green") +
  theme(plot.background = element_rect(fill = "white"),
      panel.background = element_rect(fill = "white"),        
      panel.grid.major = element_line(colour = "white",size=0.25),
      panel.grid.minor = element_blank(),
      axis.text.x=element_text(size=10,angle=90, hjust=1, 
                               face="plain", family="serif"),
      axis.text.y=element_text(size=10, hjust=1, face="plain", family="serif"), 
      axis.line.x = element_line(color="black", size = 0.50), 
      axis.line.y = element_line(color="black", size = 0.5))
}

The function takes data and column arguments. In this analysis, only the second through last columns will be used to make individual plots. So we call lapply as follows:

myplots <- lapply(colnames(DAT[2:ncol(DAT)]), make_plots, data = DAT)

myplots is now a list of ggplot objects which you can access with myplots[1], myplots[2],...or again with lapply.