Graph info

Should I use this graph?


This graph requires:

✅ a categorical variable

Description

A bar graph (or bar chart) is typically used to display counts for the discrete levels of a categorical variable, like political affiliation, hair color, or race/ethnicity (or species of penguin!).

Bar graphs can be arranged vertically or horizontally, but the length of the bar represents the ‘count’ for each category value.

In ggplot2, bar graphs can be built using geom_bar() (see also: geom_col()).

Getting set up

PACKAGES:

Install packages.

Code 
install.packages("palmerpenguins")
library(palmerpenguins) 
library(ggplot2)

DATA:

Artwork by @allison_horst

Filter the missing values from species in the palmerpenguins::penguins data and store it in penguins_bar.

Code 
penguins_bar <- palmerpenguins::penguins |> 
    dplyr::filter(!is.na(species))
glimpse(penguins_bar)
Rows: 344
Columns: 8
$ species           <fct> Adelie, Adelie, Adelie, Adelie, Adelie, Adelie, Adel…
$ island            <fct> Torgersen, Torgersen, Torgersen, Torgersen, Torgerse…
$ bill_length_mm    <dbl> 39.1, 39.5, 40.3, NA, 36.7, 39.3, 38.9, 39.2, 34.1, …
$ bill_depth_mm     <dbl> 18.7, 17.4, 18.0, NA, 19.3, 20.6, 17.8, 19.6, 18.1, …
$ flipper_length_mm <int> 181, 186, 195, NA, 193, 190, 181, 195, 193, 190, 186…
$ body_mass_g       <int> 3750, 3800, 3250, NA, 3450, 3650, 3625, 4675, 3475, …
$ sex               <fct> male, female, female, NA, female, male, female, male…
$ year              <int> 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007…

The grammar

CODE:

Create labels with labs()

Initialize the graph with ggplot() and provide data

Map species to the x axis

Map species to the fill aesthetic inside the aes() of geom_bar()

Remove the legend with show.legend = FALSE

Code 
labs_bar <- labs(
  title = "Adult foraging penguins",
  subtitle = "Distribution of flipper length",
  x = "Species", y = "Count", 
  fill = "Species")
ggp2_bar <- ggplot(data = penguins_bar,
       aes(x = species)) +
    geom_bar(aes(fill = species), 
        show.legend = FALSE)
ggp2_bar +
  labs_bar

GRAPH:

More info

  • The connection between statistical transformations and geoms is an important principle for building graphs (and mastering the grammar) with ggplot2
    • Below we cover why geom_bar(stat = "count") produces the same result as stat_count(geom = "bar")

“every geom has a default stat, and every stat a default geom.” - ggplot2 book

  • Bar graphs can also be created with geom_col()

stat_count():

The default stat argument in geom_bar() is set to "count", which ‘counts the number of cases at each x position’, so it’s ideal for categorical variables (or factors).

The stat_count() function can also be used to create bar graphs using the geom argument.

The link between geom_geom_name(stat = "stat_name") and stat_stat_name(geom = "geom_name") is shown below:

Code 
ggp2_geom_bar <- ggplot(data = penguins_bar,
       aes(x = species)) +
    geom_bar(aes(fill = species), 
        stat = "count") + 
    labs(title = "geom_bar(stat = 'count')")
ggp2_geom_bar
ggp2_stat_count <- ggplot(data = penguins_bar,
       aes(x = species)) +
    stat_count(aes(fill = species), 
        geom = "bar") + 
    labs(title = "stat_count(geom = 'bar')")
ggp2_stat_count

geom_col:

To create a bar graph with geom_col(), the count variable needs to be computed before being mapped into the graph y aesthetic.

Code 
penguins_bar |> 
    # create column of counts 
    dplyr::count(species, name = "count") |> 
    # map into x and y
    ggplot(mapping = aes(x = species, y = count)) +
    geom_col(aes(fill = species), 
        show.legend = FALSE) + 
    labs_bar +
    labs(caption = "*created with geom_col()")
# compare to geom_bar()
ggp2_bar +
  labs_bar