how to add legend in R to a ggplot

Question

I have a problem about R, something should be wrong here, I want to add a legend for the two different variables, one is "Insgesamt_ALL" with red colour,and the other one is "weiblich_ALL" with black colour.

> data<-read.csv("DatensatzUE4_01.csv",sep=";",header=T)

> nDATA<-subset(data,data$Fach=='Wirtschaftsingenieurw.m.wirtschaftswiss.Schwerpkt.')
> 
> library(ggplot2)
> 
> p <- ggplot(data=nDATA,aes(x=Semester,fill=y))
+     ggtitle("GGplot")
+     xlab("Year and Semester")
+     ylab("Total of the student")
+     geom_bar(size=3,aes(y=Insgesamt_ALL,fill=Semester),stat="identity",fill="red")
+     geom_bar(size=3,aes(y=weiblich_ALL,fill=Semester),stat="identity",fill="black")
> 
> p2<-p+ 
    theme(panel.background=element_rect(fill="white"),
          panel.grid.major=element_line(colour="grey",size=0.1),
          plot.title=element_text(face="bold"),
          axis.text=element_text(colour="black"),
          axis.title.x=element_text(face="bold"),
           axis.text.x=element_text(angle=90) )
> 
> plot(p2)

Result:

The reason for the missing legend is that you fix the filling to a constant value by using `fill="red"` and `fill="black"`. The legend is only printed automatically if you use fill as an aesthetic. You are not using `ggplot()` in the way it is intended to be used and I encourage you to learn it properly from some introduction. — Stibu, Dec 23 '14 at 20:16
Put your code in a code block using `{}`. Do *not* use `"` for that. — jlhoward, Dec 23 '14 at 20:55
Also, you should provide your data so the example is reproducible. Best way is by posting the output of `dput(data)`. If the dataset is very large, upload it somewhere and post a link. — jlhoward, Dec 23 '14 at 20:57

score 1 · Accepted Answer · answered Dec 24 '14 at 09:37

The basic issue with your missing legend is that you should take advantage of ggplot fill aesthetic by mapping it to a variable. After that you can modify the fill colors as you like. You have 2 variables: Insgesamt_ALL and weiblich_ALL.

First of all let's build some fake data (see @jlhoward comment) that mimics your actual dataset:

(tmp_data <- data.frame(Semester = seq(1:12), Insgesamt_ALL = sample(0:3000, 12),     weiblich_ALL = sample(2000:5000, 12)))

   Semester Insgesamt_ALL weiblich_ALL
1         1          2264         2643
2         2           244         3742
3         3          1681         2897
4         4          1037         4342
5         5          1225         4384
6         6           478         2195
7         7            97         2948
8         8          2537         3509
9         9          1210         3892
10       10          2016         2507
11       11          2524         2415
12       12           427         4167

First key point is that you should feed ggplot a set of key/value observations, so let's reshape the dataset:

library(tidyr)
nDATA    <- gather(tmp_data, variable, count_of_student, Insgesamt_ALL, weiblich_ALL)

Here I used tidyr::gather but any other tools would be ok as well. Let's plot it straight away:

library(ggplot2)
p <- ggplot(nDATA) + geom_bar(aes(x = Semester, y = count_of_student, fill = variable), stat = "identity")
plot(p)

enter image description here

What you are after is basically changing the fill scale to custom colors (black and red):

fill_colors        <- c("#000000", "#FF0000")
names(fill_colors) <- levels(nDATA$variable)
fill_scale <- scale_fill_manual(name = "Variable", values = fill_colors)

p + fill_scale

enter image description here

Finally, let's switch black and red fill colors by reordering the levels of the variable factor:

nDATA$variable         <- relevel(nDATA$variable, ref = "weiblich_ALL")
new_fill_colors        <- fill_colors
names(new_fill_colors) <- levels(nDATA$variable)
new_fill_scale <- scale_fill_manual(name = "Variable", values = new_fill_colors)
p + new_fill_scale

enter image description here

You should be on the right track, now.

how to add legend in R to a ggplot

1 Answers1