How to produce different geom_vline in different facets in R?

Question

I am trying to produce 2 different geom_vlines with different colours in 2 different facets of a dataset. I am doing this to highlight means of 2 different facets.

Here's the dataset:

Pclass  Sex    Age  SibSp   Parch   Fare    Cabin   Embarked    Survived
  3     male    22   1        0     7.25                S          0    
  1     female  38   1        0    71.2833   C85        C          1
  3     female  26   0        0     7.925               S          1    
  1     female  35   1        0    53.1     C123        S          1
  3     male    35   0        0     8.05                S          0    
  1     male    54   0        0    51.8625   E46        S          0

Here's the code:

g<-ggplot(data = train3, aes(x = Age, y = Survived, colour = factor(Pclass)))
g<-g+facet_wrap(~Sex)
g<-g+geom_point(size = 4, alpha = 0.2)+ggtitle("Survival by Gender")+theme(plot.title = element_text(hjust = 0.5))
g<-g+geom_vline(data = subset(train3,Sex=="female"), xintercept = mean(train3[which(train3$Sex=="female"),3]), colour = "pink", size = 1)
g<-g+geom_vline(data = subset(train3,Sex=="male"), xintercept = mean(train3[which(train3$Sex=="male"),3]), colour = "blue", size = 1)
g

Here's the output

I actually want to produce only 1 vline in each facet: pink in female and blue in male.

The suggestion given here is not working either . Error shown being:

Error in .(Sex == "female") : could not find function "."

Ptarmigan

reproducible example please. – Munt

Answer 1

Here's how you can put in different geom_vline for different iris species:

ggplot(iris, aes(Sepal.Length, Petal.Length)) + facet_wrap(~Species, scales="free") + geom_point() + 
  geom_vline(data=filter(iris, Species=="setosa"), aes(xintercept=5), colour="pink") + 
  geom_vline(data=filter(iris, Species=="versicolor"), aes(xintercept=6), colour="blue") + 
  geom_hline(data=filter(iris, Species=="virginica"), aes(yintercept=6), colour="green") 

Munt

Legend!! apparently that subset function in data call wasn't working. Glad to find that filter worked. Thanks for the quick help. Much appreciated. – Ptarmigan

I thought the error was due to the xintercept not stated in aes? – Munt

Compare geom_vline(data=subset(iris, Species=="setosa"), aes(xintercept=5), colour="pink") versus geom_vline(data=subset(iris, Species=="setosa"), xintercept=5, colour="pink") – Munt

This works fine for 3 facets, but is there any way to do this efficiently for large numbers of facets? – Marche

This looks like it would work for multiple facets more easily #10942860 – Calamite

Answer 2

You can create a data.frame with one column being intercept values to be used for lines and a second column with Sex. So that when using facet_wrap, they are separated.
Something like:

dataInt <- train3 %>%
  group_by(Sex) %>%
  summarize(Int = mean(Age))

Then you can use it in your script:

g<-ggplot(data = train3, aes(x = Age, y = Survived, colour = factor(Pclass))) + 
  facet_wrap(~Sex) +
  geom_vline(data=dataInt, aes(xintercept=Int))

Without your data, I cannot test this.

[EDIT: With a reprex] If I use the same data as Adam Quek for a reproducible example, the code would be automated as follows:

library(tidyverse)

dataLine <- iris %>%
  group_by(Species) %>%
  summarize(mean_x = mean(Sepal.Length))

ggplot(iris) +
  aes(x = Sepal.Length, y = Petal.Length) +
  facet_wrap(facets = vars(Species)) + 
  geom_point() + 
  geom_vline(data  = dataLine, aes(xintercept = mean_x, colour = Species))

Shephard

Rather than creating a new data set dataInt with the function summarize(Int = mean(Age)), which didn't work for me, use mutate instead. – Cohesion

This answer is great as it allows the lines to be flexible based on automatically computed values from the data. – Pily

FYI if interested in adjusting properties of the line use syntax like the following: + geom_vline(data = dataLine, aes(xintercept = mean_x), linetype="dashed", size = 1.5, colour = "green") – Pily

Answer 3

Building on @Sébastien Rochette's answer above; Rather than creating a new data frame dataInt with the function summarize(Int = mean(Age)), which didn't work for me as I had multiple levels within each facet plot, use mutate instead.

train3 <- train3 %>%
  group_by(Sex) %>%
  mutate(Int = mean(Age))

And then you can use train3 data-frame in

g<-ggplot(data = train3, aes(x = Age, y = Survived, colour = factor(Pclass))) + 
  facet_wrap(~Sex) +
  geom_vline(data=train3, xintercept=Int)

This works but I fear it may have created a geom_vline for all values, because each mean will be repeated within each level of each factor within the dataframe.

Cohesion