I want to create a plot with the x-axis at the top and the axis labels rotated 90 degrees center-aligned with the ticks. This is my code:
library(tidyverse)
# create data
a3 <- rnorm(100, 0, 1)
aaa4 <- rnorm(100, 0, 1)
aaaa5 <- rnorm(100, 0, 1)
df_2 <- data.frame(a3, aaa4, aaaa5)
correlation_matrix_2 <- cor(df_2)
corr_long <- correlation_matrix_2 %>%
as.data.frame(.) %>%
rownames_to_column(., var = "V1") %>%
pivot_longer(., cols = -V1, names_to = "V2", values_to = "r")
# plot:
ggplot(corr_long, aes(V1, V2))+
geom_tile(mapping = aes(fill = r))+
scale_x_discrete(position = "top")+
theme(axis.text.x = element_text(size = 16, angle = 90, hjust = 0, vjust = .5))
However, the plot ignores the vjust parameter of the theme and always positions the labels a little bit off. This only happens when the axis is set to be at the top using position parameter from scale_x_discrete, it works fine for "bottom". The hjust works as expected in both cases.
I then tried a few other things and found that it also happens when the y axis is set to the right, but then it is hjust that is ignored and vjust works. It also happens for scale_x/y_continuous. Changing the order (theme first, then scale_x_discrete) does not change anything either and also splitting it in two separate commands does not work:
[...]
theme(axis.text.x = element_text(hjust = 0))+
theme(axis.text.x = element_text(vjust = 1))+
[...]
Is that a bug or do I miss something here?
You can adjust the position of the axis text at the top using the axis.text.x.top argument of the theme() function instead of axis.text.x. Similarly, you can use axis.text.y.right for y-axis text on the right hand side. For example:
ggplot(corr_long, aes(V1, V2))+
geom_tile(mapping = aes(fill = r))+
scale_x_discrete(position = "top")+
theme(axis.text.x.top = element_text(size = 16, angle = 90, hjust = 0, vjust = .5))