How can I add second x axis labels to a bar plot?

Question

I am trying to produce a graph similar to the one below: graph with two x axis labels, facet wrap and different colours

The format of my data is as follows:Data format

The code that I have been using is:

data$x1 <- as.factor(data$x1)
data$x2 <- as.factor(data$x2)
data$Fill <- as.factor(data$Fill)

ggplot(data, aes(x=x1, y=y, fill=x2))+
  geom_col(position=position_dodge())+
  theme( legend.position = "right")+
  facet_wrap(~Sex, nrow=1, strip.position = "top")+
  theme(strip.placement = "outside")
  scale_x_discrete(breaks = c(11, 12, 22),labels = c("11", "12", "22"))+
  ylab("y") 

This produces the following plot:Plot from above code

However, I am unable to add x2 as a second variable on the x axis nor am I able to then use Fill as the fill= option. Does anyone have a solution?

Kind regards

I have googled the issue and played around with the code but without success.

Answer 1

Instead of using position_dodge one approach would be to create a new column by pasting your x1 and x2 columns together separated by a line break and mapping this new column on x. This way you will automatically get your two rows of axis labels. Tricky first part is to put the labels in the right order for which I use arrange to order the dataset first and second use forcats::fct_inorder to convert to a factor and setting the right order. A drawback of this approach is that the space between each bar is the same and groups of bars are no longer separated.

And as you already facetted by Sex there is no easy fix for that. One approach would be to use facet_grid or as I do below, to use ggh4x::facet_nested_wrap to facet by both Sex and x1 then get rid of the strips for x1. Drawback of this approach is that it leaves some blank space.

Using some fake example data:

library(dplyr, warn = FALSE)
library(ggplot2)
library(ggh4x)

# Example data
data <- expand.grid(
  Sex = 1:2,
  x1 = c("11", "12", "22"),
  x2 = c("11", "12", "22")
) |>
  dplyr::mutate(
    Fill = purrr::map2_chr(x1, x2, \(x, y) paste0(sort(c(x, y)), collapse = "")),
    Fill = as.numeric(factor(Fill)) - 1,
    y = seq_len(dplyr::n())
  )

data$x1 <- as.factor(data$x1)
data$x2 <- as.factor(data$x2)
data$Fill <- as.factor(data$Fill)

data <- data |>
  arrange(Sex, x1, x2) |>
  mutate(
    x = paste(x2, x1, sep = "\n"),
    x = forcats::fct_inorder(x)
  )

ggplot(data, aes(x = x, y = y, fill = Fill)) +
  geom_col() +
  facet_nested_wrap(~ Sex + x1,
    nrow = 1,
    scales = "free_x",
    strip = ggh4x::strip_nested(
      by_layer_x = TRUE,
      text_x = list(element_text(), element_blank()),
      background_x = list(element_rect(), element_blank())
    )
  ) +
  scale_x_discrete(expand = c(0, .8))