Mastering Bar Charts in ggplot2

Introduction

Bar charts are essential for comparing the frequency or count of categorical variables. In this tutorial, we will use the starwars dataset to build a professional bar chart layer by layer.

1. Environment & Typography Setup

First, we load the necessary libraries and configure the Poppins Google font to give our plots a modern, professional look.

Code

library(tidyverse)
library(showtext)
library(sysfonts)

# Add the Google font "Poppins"
font_add_google("Poppins", "poppins")
showtext_auto()

# Set Poppins as the default font for all plots
theme_set(theme_test(base_size = 15) + 
            theme(text = element_text(family = "poppins")))

2. Basic Bar Chart

We begin by mapping the sex variable to the X-axis. geom_bar() will automatically count the observations for each category.

Code

starwars %>% 
  ggplot(aes(x = sex)) +
  geom_bar()

3. Aesthetic Refinement

To make the plot visually appealing, we add color and adjust the transparency.

Code

starwars %>% 
  ggplot(aes(x = sex)) +
  geom_bar(fill = "purple", alpha = 0.6) +
  theme_test() +
  labs(title = "Distribution of Sex in Starwars",
       subtitle = "Basic aesthetic customization",
       x = "Sex",
       y = "Count")

4. Systemic Data Cleaning

Missing values (NAs) can clutter a graph. We use drop_na() to target specific columns for cleaning.

Code

starwars %>% 
  # Precise cleaning: only remove rows where sex is missing
  drop_na(sex) %>% 
  ggplot(aes(x = sex)) +
  geom_bar(fill = "purple", alpha = 0.6) +
  theme_test() +
  labs(title = "Cleaned Bar Chart",
       subtitle = "Removing NA values from the sex variable",
       x = "Sex",
       y = "Count")

5. Improving Readability (Flip & Sort)

Coordinate Flip

Flipping the chart is useful when you have many categories or long labels.

Code

starwars %>% 
  drop_na(sex) %>% 
  ggplot(aes(x = sex)) +
  geom_bar(fill = "purple", alpha = 0.6) +
  coord_flip() +
  theme_test() +
  labs(title = "Horizontal Bar Chart",
       x = "Sex",
       y = "Count")

Sorting by Frequency

The most professional way to present a bar chart is by sorting the categories from most frequent to least frequent.

Code

starwars %>% 
  drop_na(sex) %>% 
  ggplot(aes(x = fct_infreq(sex))) +
  geom_bar(fill = "purple", alpha = 0.6) +
  theme_test() +
  labs(title = "Frequency Sorted Bar Chart",
       subtitle = "Ordered from highest to lowest frequency",
       x = "Biological Sex",
       y = "Total Characters",
       caption = "Data: starwars dataset | Prepared by Abdullah Al Shamim")

Systemic Summary Toolkit

Function	Role	Systemic Importance
`geom_bar()`	Statistical Geometry	Automatically calculates counts.
`drop_na(col)`	Targeted Cleaning	Ensures data integrity without over-filtering.
`fct_infreq()`	Factor Reordering	Provides instant visual hierarchy.
`coord_flip()`	Spatial Adjustment	Optimizes plot for readability.