I have adata.frame such as:

df <- data.frame(region = c(rep("r1", 5), rep("r2", 5)),
                 family = c("A1", "A1", "A2", "A2", "A3",
                            "B1", "B6", "B11", "B2", "B2"),
                 genus = c("a1", "a2", "a1", "a1", "a1",
                           "b1", "b3", "b3", "b1", "b4"),
                 spp = c("a1a", "a4a", "a2a", "a1a", "a1a",
                         "b1b", "b6b", "b6b", "b2b", "b3b"))

I'm trying to create a grouped bar chart such as the following image

Where each color is respectively family, genus, and spp and "R1"/"R2 are region. The N axis must be the n_distinct() of family, genus, and spp. How can I approach this?

CodePudding user response：

Get your data into shape:

library(ggplot2)
library(tidyr)

plot_df  <- df  |>
    group_by(region)  |>
    summarise(across(family:spp, n_distinct))  |>
    pivot_longer(-region)

plot_df
# A tibble: 6 x 3    
#   region name   value
#   <chr>  <chr>  <int>
# 1 r1     family     3
# 2 r1     genus      2
# 3 r1     spp        3
# 4 r2     family     4
# 5 r2     genus      3
# 6 r2     spp        4

Then you can plot it:

ggplot(plot_df)   
    geom_col(mapping = aes(
        x = name,
        y = value,
        fill = name
    ))  
    facet_wrap(vars(region))  
    labs(
        y = "n",
        title = "Number of distinct whatever"
    )  
    theme_bw()  
    theme(
        axis.title.x = element_blank(),
        legend.title = element_blank()
    )  
    scale_fill_brewer(palette=3)

Edit: made the plot look slightly nicer.

CodePudding user response：

You can plot region directly on the x axis without faceting too:

library(tidyverse)

pivot_longer(df, family:spp, names_to = 'level') %>%
  group_by(region, level) %>% 
  summarize(n = n_distinct(value)) %>%
  ggplot(aes(region, n, color = level))  
  geom_col(position = position_dodge(width = 0.7), width = 0.5,
           fill = 'white', size = 3)  
  theme_classic(base_size = 16)  
  scale_color_manual(values = c('#ff7f27', '#a349a4', '#ffaec9'))