Adding a legend to plot

Question

I can't seem to get a legend to appear on this plot. Forums and web tutorials have not provided a solution. Any help would be appreciated.

I have created a bar plot that overlays frequencies when NA values are included and excluded but I need to provide a legend for the plot.

Full_Freq<-data.frame(NA_INC = c(11.8, 10.1, 11.9, 18.8, 17.8, 1.7, 1.7, 1.8, 3.5, .001),
                      NA_EXC = c(54.9, 38.6, 27.9, 34.7, 36.1, 2.3, 2.3, 2.8, 8.4, 100.0),
                      code = c("RI1 - R1,C1,A,G,R", "RI2 - R1,C1,A,E,R", "RC1 - R1,C1,A,E,R",
                               "RP3 - R2,C2,B1L1,BGLG,BRLR", "RP4 - R1,C1,A,P,O", "RM1 A - R3,C3,M1D1L1,MWDELE,MRDRLR",
                               "RM1 B - R3,C4,M2D1L1,MWDPLP,MRDRLO","RM2 - R3,C3,M1D1L1,MWDGLG,MRDRLR",
                               "RM3 - R3,C3,M1D1L1,MWDGLG,MRDRLR","RM4 - R1,C1,A,P,R"))

ggplot(Full_Freq) + 
  geom_col(aes(x = code, y = NA_EXC), fill = "gray",  colour = "black") + 
  geom_col(aes(x = code, y = NA_INC), fill = "black",  colour = "black")+
  ylab("Percentage")+
  xlab("")+
  ggtitle("Tooth and Full Phenotype Code")+
  theme_classic()+
  theme(plot.title = element_text(hjust= 0.5))+
  theme(axis.text.x = element_text(angle = 90, vjust=0.5))

You need to reshape your data from wide to long. Gather columns `NA_INC` and `NA_EXC` then map the resulting value column to `y` and the resulting `key` column to the `fill` aesthetic. Both inside `aes()`. Related: [R - ggplot line color (using geom_line) doesn't change](https://stackoverflow.com/questions/48051905/r-ggplot-line-color-using-geom-line-doesnt-change) — markus, Apr 07 '19 at 11:01
have the answers you have gotten been helpful? If any solved your problem consider accepting one as the answer. — Kresten, May 21 '19 at 07:24

Evert · Accepted Answer · 2019-04-07T11:12:14.913

1

First transform your data in a tidy way (one observation per row):

library(tidyverse)
Full_Freq <- Full_Freq %>%
  gather(key = IncExc, value = x, -code)

Then plot,

ggplot(Full_Freq) + 
  geom_col(aes(x = code, y = x, fill = IncExc)) + 
  ylab("Percentage")+
  xlab("")+
  ggtitle("Tooth and Full Phenotype Code")+
  theme_classic()+
  theme(plot.title = element_text(hjust= 0.5))+
  theme(axis.text.x = element_text(angle = 90, vjust=0.5)) + 
  theme(legend.position = "right")

You might need to change the colors.

edited Apr 07 '19 at 11:12

answered Apr 07 '19 at 11:07

Evert

51
4

Please include all packages that you're using – pogibas Apr 07 '19 at 11:07
Why 3 calls to theme? You can put everything in the same call. – Rui Barradas Apr 07 '19 at 11:16
Thank you for your help. While this does create a legend, it alters the data and doesn't present it accurately. For instance, when RI1 has NA values included, the bar is at 11.8%, and when excluded it is at 54.9%. However, after using reshape the values add together to equal 66.7% – dalek_fred Apr 07 '19 at 18:19

Oka · Answer 2 · 2019-04-08T11:18:45.440

In many cases, making plots is easier with long data format. You can transform your data into long format with reshape2 (or also with tidyverse, as suggested in another answer):

library(reshape2)
# your code here

# reshaping your data into long format
Full_Freq.m <- melt(Full_Freq,id.vars = "code") 
Full_Freq.m
#                                 code variable   value
#1                   RI1 - R1,C1,A,G,R   NA_INC  11.800
#2                   RI2 - R1,C1,A,E,R   NA_INC  10.100
#3                   RC1 - R1,C1,A,E,R   NA_INC  11.900
#...
# 11                  RI1 - R1,C1,A,G,R   NA_EXC  54.900
# 12                  RI2 - R1,C1,A,E,R   NA_EXC  38.600
# 13                  RC1 - R1,C1,A,E,R   NA_EXC  27.900
# 14         RP3 - R2,C2,B1L1,BGLG,BRLR   NA_EXC  34.700
#...

Then making plots is easier. Plot with stacked bars:

# Making the plot
ggplot(data=Full_Freq.m, aes(x=code, y=value, fill=variable)) + geom_bar(stat="identity")+ 
  ylab("Percentage")+
  xlab("") +
  ggtitle("Tooth and Full Phenotype Code") +
  theme_classic() +
  theme(plot.title = element_text(hjust= 0.5))+
  theme(axis.text.x = element_text(angle = 90, vjust=0.5)) +   
  scale_fill_manual(values=c("grey", "black"))

Plot with dodged bars:

# dodged bar plot, e.g. bars next to each other
ggplot(data=Full_Freq.m, aes(x=code, y=value, fill=variable)) + geom_bar(stat="identity", position="dodge")+ 
  ylab("Percentage")+
  xlab("") +
  ggtitle("Tooth and Full Phenotype Code") +
  theme_classic() +
  theme(plot.title = element_text(hjust= 0.5))+
  theme(axis.text.x = element_text(angle = 90, vjust=0.5)) +   
  scale_fill_manual(values=c("grey", "black"))

Plot with overlaid bars:

# bars overlaid
ggplot(data=Full_Freq.m, aes(x=code, y=value, fill=variable)) + geom_bar(stat="identity", position="identity")+ 
  ylab("Percentage")+
  xlab("") +
  ggtitle("Tooth and Full Phenotype Code") +
  theme_classic() +
  theme(plot.title = element_text(hjust= 0.5))+
  theme(axis.text.x = element_text(angle = 90, vjust=0.5)) +   
  scale_fill_manual(values=c('NA_EXC' = 'gray', 'NA_INC' = 'black')) +
  geom_col(data=Full_Freq.m[which(Full_Freq.m$variable=="NA_INC"),])

Thank you for your help. While this does create a legend, it alters the data and doesn't present it accurately. For instance, when RI1 has NA values included, the bar is at 11.8%, and when excluded it is at 54.9%. However, after using reshape the values add together to equal 66.7% — dalek_fred, Apr 07 '19 at 18:19
And how much you think 54.9+11.8 should be? 66.7% sound about correct.. — Oka, Apr 07 '19 at 18:32
You're correct but that's not how I'm trying to visualize the data. I want the bars to overlay one another, not be stacked on top of another. — dalek_fred, Apr 07 '19 at 20:23
Ok, so you wanted a bar plot that stacks frequencies, but which is not the stacked bar plot. Now it makes sense. — Oka, Apr 07 '19 at 20:30
Yeah, I probably worded my question poorly. I did post a solution to my question though. — dalek_fred, Apr 07 '19 at 21:02
@Oka if you want bars to be placed next to each other then put the argument `position = "dodge"`in the `geom_bar()` layer — Kresten, Apr 08 '19 at 10:22
@JG1981: NP. I added that option to my answer now so it covers both, the question, and what you actually needed. — Oka, Apr 08 '19 at 11:20

score 0 · Answer 3 · answered Apr 07 '19 at 20:28

Thank you to everyone that helped. I've got my code sorted out minus having the variables on the x-axis in order. Here's how I did it -

````
### Barplot of full phenotype codes that have contrast frequencies when NA values are included/excluded 
Full_Freq<-data.frame(NA_INC = c(11.8, 10.1, 11.9, 18.8, 17.8, 1.7, 1.7, 1.8, 3.5, 0.001),
                      NA_EXC = c(54.9, 38.6, 27.9, 34.7, 36.1, 2.3, 2.3, 2.8, 8.4, 100.0),
                      code = c("RI1 - R1,C1,A,G,R", "RI2 - R1,C1,A,E,R", "RC1 - R1,C1,A,E,R",
                               "RP3 - R2,C2,B1L1,BGLG,BRLR", "RP4 - R1,C1,A,P,O", "RM1 A - R3,C3,M1D1L1,MWDELE,MRDRLR",
                               "RM1 B - R3,C4,M2D1L1,MWDPLP,MRDRLO","RM2 - R3,C3,M1D1L1,MWDGLG,MRDRLR",
                               "RM3 - R3,C3,M1D1L1,MWDGLG,MRDRLR","RM4 - R1,C1,A,P,R"))


#### make the plot
ggplot(Full_Freq) + 
  geom_col(aes(x = code, y = NA_EXC, fill = "NA excluded")) + 
  geom_col(aes(x = code, y = NA_INC, fill= "NA included")) +
  scale_fill_manual(name = 'NA Values', values = c('NA excluded' = 'gray', 'NA included' = 'black')) +
  ylab("Percentage")+
  xlab("")+
  ggtitle("Full Phenotype Code for Maxillary Teeth")+
  theme_classic()+
  theme(plot.title = element_text(hjust= 0.5))+
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust=0.5))
````

Kresten · Answer 4 · 2019-04-09T06:44:55.823

Adding to @Oka solution:

tidy your data. I use tidyr::gather()
plot dodged bar plot
- reorder the x axis using reorder()
fix theme, add lables

code

Full_Freq %>% 
tidyr::gather("variable", "value", contains("NA")) %>% 
ggplot(aes(x=reorder(code, value, max), y=value, fill=variable)) + geom_bar(stat="identity", position="dodge")+ 
  labs(fill = "legend title",
       x= "x title",
       y = "Percentage",
       title = "Tooth and Full Phenotype Code")+
  theme_classic() +
  theme(axis.text.x = element_text(angle = 90, vjust=0.5)) +   
  scale_fill_manual(values=c("grey", "black"))

Adding a legend to plot

4 Answers4