Bar Charts

This lesson is called Bar Charts, part of the Fundamentals of R course. This lesson is called Bar Charts, part of the Fundamentals of R course.

Transcript

Click on the transcript to go to that point in the video. Please note that transcripts are auto generated and may contain minor inaccuracies.

View code shown in video

# Load Packages -----------------------------------------------------------

library(tidyverse)

# Import Data -------------------------------------------------------------

penguins <- read_csv("penguins.csv")
			
# Bar Charts --------------------------------------------------------------

# There are two basic approaches to making bar charts, 
# both of which use geom_bar().

# Approach #1

# Use your full dataset.
# Only assign a variable to the x axis.
# Let ggplot use the default stat transformation (stat = "count") 
# to generate counts that it then plots on the y axis.

ggplot(data = penguins,
       mapping = aes(x = bill_length_mm)) +
  geom_bar()

# The default statistical transformation for geom_bar() is count. 
# This will give us the same result as our previous plot.

ggplot(data = penguins,
       mapping = aes(x = bill_length_mm)) +
  geom_bar(stat = "count")


# Approach #2

# Wrangle your data frame before plotting, creating a new data frame 
# in the process
# Assign variables to the x and y axes
# Use stat = "identity" to tell ggplot to use the data exactly as it is

# It's often easier to do our analysis work, save a data frame, 
# and then use this to plot.
# Let's recreate our penguin_bill_length_by_island data frame.

penguin_bill_length_by_island <- penguins |> 
  group_by(island) |> 
  summarize(mean_bill_length = mean(bill_length_mm, na.rm = TRUE)) |> 
  arrange(mean_bill_length)

# Then let's use this data frame to make a bar chart. 
# The stat = "identity" here tells ggplot to use the exact data points
# without any statistical transformations.

ggplot(data = penguin_bill_length_by_island,
       mapping = aes(x = island,
                     y = mean_bill_length)) +
  geom_bar(stat = "identity")

# We can also flip the x and y axes. 

ggplot(data = penguin_bill_length_by_island,
       mapping = aes(x = mean_bill_length,
                     y = island)) +
  geom_bar(stat = "identity") 

# The function coord_flip() will do the same thing.

ggplot(data = penguin_bill_length_by_island,
       mapping = aes(x = island,
                     y = mean_bill_length)) +
  geom_bar(stat = "identity") +
  coord_flip()

# We can also use geom_col(), which is the same as geom_bar(stat = "identity")

ggplot(data = penguin_bill_length_by_island,
       mapping = aes(x = island,
                     y = mean_bill_length)) +
  geom_col()

Your Turn

# Load Packages -----------------------------------------------------------

library(tidyverse)

# Import Data -------------------------------------------------------------

penguins <- read_csv("penguins.csv")
			
# Bar Charts --------------------------------------------------------------

# Use the v1 approach to make a bar chart that shows a count of the number of penguins by species. 

# YOUR CODE HERE

# Use the v2 approach by doing the following:
# 1. Creating a new data frame called penguins_by_species that is a 
# count of the number of penguins by species
# 2. Plot your data frame using the v2 approach with geom_bar()

# YOUR CODE HERE

# Make the same graph as above, but use geom_col() instead of geom_bar()

# YOUR CODE HERE

Learn More

You can also find examples of code to make bar charts on the Data to Viz website , the R Graph Gallery website , and in Chapter 3 of the R Graphics Cookbook. Michael Toth also has a detailed blog post about making bar charts with ggplot.

Have any questions? Put them below and we will help you out!

You need to be signed-in to comment on this post. Login.

John LeMay • October 18, 2024

Why do the extra data wrangling for v2? Is it in case you want to do something besides a basic count for each bar? I am a bit confused on that.

David Keyes Founder • October 19, 2024

Good question! My experience is that you only rarely want to make a simple enough plot that the v1 approach works. Usually, I'm doing a bunch of filtering, summarizing, etc before I plot. In that case, it's easiest (at least for how my mind works) to do this, save it as an object, and then use this to plot. Hope that helps!