make grouping bar plot ggplot2

Question

I'm new to ggplot2 but trying to use it. I have two variables: SA( with 4 levels :0, 1000,2000 and 3000) and GA (with 4 levels:0, 0.5,1 and 2). I would like to group these by SA (like this Figure)

> G<- read.table("k.csv", sep=";",header = TRUE)
> G
SA  GA PH
1     0 0.0 41
2     0 0.0 27
3     0 0.0 28
4     0 0.0 25
5     0 0.5 35
6     0 0.5 45
7     0 0.5 35
8     0 0.5 55
9     0 1.0 45
10    0 1.0 35
11    0 1.0 38
12    0 1.0 46
13    0 2.0 52
14    0 2.0 40
15    0 2.0 40
16    0 2.0 35
17 1000 0.0 30
18 1000 0.0 30
19 1000 0.0 30
20 1000 0.0 30
21 1000 0.5 28
22 1000 0.5 33
23 1000 0.5 31
24 1000 0.5 42
25 1000 1.0 38
26 1000 1.0 30
27 1000 1.0 27
28 1000 1.0 25
29 1000 2.0 30
30 1000 2.0 22
31 1000 2.0 31
32 1000 2.0 44
33 2000 0.0 18
34 2000 0.0 25
35 2000 0.0 24
36 2000 0.0 31
37 2000 0.5 24
38 2000 0.5 22
39 2000 0.5 36
40 2000 0.5 40
41 2000 1.0 27
42 2000 1.0 29
43 2000 1.0 42
44 2000 1.0 33
45 2000 2.0 20
46 2000 2.0 40
47 2000 2.0 30
48 2000 2.0 25
49 3000 0.0  0
50 3000 0.0  0
51 3000 0.0  0
52 3000 0.0  0
53 3000 0.5 24
54 3000 0.5 20
55 3000 0.5 25
56 3000 0.5 NA
57 3000 1.0 37
58 3000 1.0 NA
59 3000 1.0 38
60 3000 1.0 25
61 3000 2.0 24
62 3000 2.0 15
63 3000 2.0 20
64 3000 2.0 32
> ggplot(G, aes(x=SA, y=PH, fill=factor(GA))) +   
stat_summary(geom="bar",positiGon=position_dodge(1))

but it does not give me what I need. It gives me something different (here it is)

Also, I would like to add error bar to the bars. Any ideas?

pogibas · Accepted Answer · 2017-10-22 07:38:33Z

2

Solution where I use data.table to calculate standard error and mean.

library(data.table)
library(ggplot2)
setDT(G)
pd <- G[, .(SE = sd(PH, na.rm = TRUE) / sqrt(.N),
            MN = mean(PH, na.rm = TRUE)), 
           .(SA, GA)]
ggplot(pd, aes(factor(SA), fill = factor(GA))) + 
    geom_bar(aes(y = MN), 
             stat = "identity", position = "dodge") + 
    geom_errorbar(aes(ymin = MN - SE, ymax = MN + SE),
                  position = "dodge") +
    labs(x = "SA",
         y = "PH",
         fill = "GA") +
    theme_classic()

answered Oct 22, 2017 at 7:38

pogibas

28.5k21 gold badges92 silver badges120 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Adam Quek · Accepted Answer · 2017-10-22 00:16:30Z

1

library(tidyr)

std.error <- function(x) sd(x)/sqrt(length(x))
means <- function(x)mean(x, na.rm=TRUE)


df2 <- G %>% 
       group_by(SA, GA) %>% 
       mutate(error=std.error(PH)) %>% 
       summarise_at(vars(PH:error), funs(means))

ggplot(df2, aes(as.factor(SA), PH, fill=as.factor(GA))) + 
    geom_bar(stat="identity", position="dodge") + 
    geom_errorbar(aes(ymin=PH-error, ymax=PH+error), 
                  width=.2, position=position_dodge(.9))

edited Oct 22, 2017 at 0:16

answered Oct 21, 2017 at 23:42

Adam Quek

7,1831 gold badge20 silver badges24 bronze badges

4 Comments

user3801226 Over a year ago

thanks Adam ! however, it gives me an error with Error in group_by(., SA, GA) : could not find function "group_by"

tyluRp Over a year ago

group_by is a function from the dplyr package. You will have to load this package in order to use group_by.

Adam Quek Over a year ago

try installing and loading the packages of "tidyverse", i.e. install.packages("tidyverse"); library(tidyverse)

user3801226 Over a year ago

I have installed and loaded both dplyr and tidyverse but it is still the same error message. > library(tidyverse) Error: package or namespace load failed for ‘tidyverse’ in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]]): there is no package called ‘bindr’ In addition: Warning message: package ‘tidyverse’ was built under R version 3.4.2

Collectives™ on Stack Overflow

make grouping bar plot ggplot2

2 Answers 2

Comments

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related