Changing of color for geom_text gives completely different color then called for

Question

I want to add labels to my ggplot2 bars and change the color of the label. Somehow I can't.

My dataset is (simplified) approximatly in this format:

data$value <- runif(27, min=10, max=60)
data$A <- factor((rep(1:3, each=9)))
data$B <- factor((rep(1:9, each=3)))
data$variable <- factor(rep(rep(1:3),9))

The plot would be like:

three <- c(pink="#BD1550",dark="#490A3D",blue1="#0b6fa1",white="#FFFFFF", "#FFFFFF")
 m<-  data %>% group_by(A, variable) %>% summarise(mean=mean(value), sd=sd(value)) %>% 
             ggplot(aes(x=A,fill=variable)) +
             geom_col(aes(y=mean),position="stack")+
             geom_text(aes(label=round(mean,digits=2),y=mean, colour="white")
                          ,size=3, show.legend = F, position = position_stack(vjust = 0.5))+
             scale_fill_manual(values=three) + theme(legend.position="right")

Now, for the colour in geom_text I have tried:

1. color="white"
2. spelling color or colour
3. colour="‎#FFFFFF"
4. colour=c("#FFFFFF")
5. colour = 4
6. color = white
7. one <- c("#FFFFFF") and then colour = one

Different solutions give different colors for each label, pink, orange, green, a blue from my string 'three', but never give me the color white. I have also tried to make it different colors than white, but somehow I have no control over what color it gives me back.

I get no error messages.

I am starting to run out of ideas. Anyone any solutions?

Answer 1

The problem is that your are mapping "white" on the color aesthetic inside aes(). This way ggplot thinks you want to map a variable on the color aesthetic, i.e. "white" is not interpreted as the name of a color. Instead ggplot simply picks the color from its default palette, which is "red". Simply pass the color as an argument to geom_text outside of aes(). Or use scale_color_manual to set the color palette. (; Try this:

library(ggplot2)
library(dplyr)

set.seed(42)

data <- data.frame(
  value = runif(27, min=10, max=60),
  A = factor((rep(1:3, each=9))),
  B = factor((rep(1:9, each=3))),
  variable = factor(rep(rep(1:3),9))  
)

three <- c(pink="#BD1550",dark="#490A3D",blue1="#0b6fa1", white="#FFFFFF", "#FFFFFF")
m <-  data %>% 
  group_by(A, variable) %>% 
  summarise(mean=mean(value), sd=sd(value)) %>% 
  ggplot(aes(x=A, fill=variable)) +
  geom_col(aes(y = mean),position="stack")+
  geom_text(aes(label = round(mean, digits=2), y=mean), colour="white"
            ,size=3, show.legend = F, position = position_stack(vjust = 0.5))+
  scale_fill_manual(values=three) + theme(legend.position="right")

m

^{Created on 2020-04-14 by the reprex package (v0.3.0)}