How to Show a Hierarchical Structure on the Axis Labels with ggplot2 in R?

Visualizing data with a hierarchical structure can be a bit challenging, especially when you want to convey multiple levels of grouping or categorization within a single plot. In R, using ggplot2, we can effectively represent this hierarchical information in the axis labels, making it easier to understand the structure and relationships between different levels of data. This article will guide you through how to display hierarchical structure on the axis labels using ggplot2 in R.

What is a Hierarchical Structure?

A hierarchical structure represents data in multiple layers or levels, such as categories and subcategories, showing how data points are organized. For example, you might have sales data for different regions, and within each region, you have different product categories.

Why Display Hierarchical Structures on Axis Labels?

• Better Data Interpretation: Displaying hierarchical information helps the viewer understand the relationships between different categories.
• Detailed Analysis: It allows you to present multiple levels of grouping within a single visualization, making complex datasets more interpretable.
• Clear Presentation: A well-structured plot with hierarchical labels can effectively communicate insights and trends across different levels of data.

We will be using a combination of data manipulation and plotting techniques to display these hierarchical structures using R Programming Language.

1. Installing and Loading Required Packages

To get started, install and load the necessary packages. We'll use ggplot2 for plotting and dplyr for data manipulation.

# Install required packages if not already installed
install.packages("ggplot2")
install.packages("dplyr")

# Load the required libraries
library(ggplot2)
library(dplyr)

2. Creating a Sample Hierarchical Dataset

Let’s create a simple dataset representing sales data across multiple regions and product categories.

# Creating a sample dataset
sales_data <- data.frame(
  Region = c("North", "North", "North", "South", "South", "South", "East", "East", "East"),
  Category = c("Electronics", "Furniture", "Clothing", "Electronics", "Furniture", "Clothing", 
               "Electronics", "Furniture", "Clothing"),
  Sales = c(15000, 12000, 8000, 17000, 13000, 9000, 16000, 14000, 8500)
)

# Display the dataset
print(sales_data)

Output:

  Region    Category Sales
1  North Electronics 15000
2  North   Furniture 12000
3  North    Clothing  8000
4  South Electronics 17000
5  South   Furniture 13000
6  South    Clothing  9000
7   East Electronics 16000
8   East   Furniture 14000
9   East    Clothing  8500

• Region: The main grouping (North, South, East)
• Category: Sub-grouping (Electronics, Furniture, Clothing)
• Sales: The sales figures

3. Combining Hierarchical Levels into One Label

To display hierarchical data on the X-axis, we need to combine the Region and Category columns into a single label. We'll use the paste() function to do this.

# Combine Region and Category columns to create a hierarchical label
sales_data <- sales_data %>%
  mutate(Hierarchical_Label = paste(Region, Category, sep = "\n"))

# Display the updated dataset
print(sales_data)

Output:

  Region    Category Sales Hierarchical_Label
1  North Electronics 15000 North\nElectronics
2  North   Furniture 12000   North\nFurniture
3  North    Clothing  8000    North\nClothing
4  South Electronics 17000 South\nElectronics
5  South   Furniture 13000   South\nFurniture
6  South    Clothing  9000    South\nClothing
7   East Electronics 16000  East\nElectronics
8   East   Furniture 14000    East\nFurniture
9   East    Clothing  8500     East\nClothing

The Hierarchical_Label column now contains values like "North\nElectronics", which represent the hierarchical structure.

4. Plotting the Data with ggplot2

Now that our dataset is ready, let's create a bar plot using ggplot2 and use our combined label as the X-axis.

# Create a bar plot with hierarchical axis labels
ggplot(sales_data, aes(x = Hierarchical_Label, y = Sales, fill = Category)) +
  geom_bar(stat = "identity") +
  labs(title = "Sales by Region and Category",
       x = "Region and Category",
       y = "Total Sales") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 10, face = "bold"))

Output:

• aes(x = Hierarchical_Label, y = Sales, fill = Category): Specifies the aesthetic mapping, where Hierarchical_Label represents the combined axis label.
• geom_bar(stat = "identity"): Creates a bar plot with the actual sales values.
• theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 10, face = "bold")): Customizes the X-axis text appearance, tilting it for better readability.

Using ggplot2’s built-in facet_grid() for More Complex Structures

For datasets with more complex hierarchical structures, you might want to use facet_grid() to split both axes.

# Creating another hierarchical dataset with additional levels
extended_sales_data <- data.frame(
  Region = rep(c("North", "South", "East"), each = 6),
  Category = rep(c("Electronics", "Furniture", "Clothing"), 6),
  Subcategory = rep(c("Laptops", "Desktops", "Chairs", "Tables", "T-Shirts", "Jeans"), 3),
  Sales = sample(10000:20000, 18, replace = TRUE)
)

# Create a plot with facet_grid
ggplot(extended_sales_data, aes(x = Subcategory, y = Sales, fill = Category)) +
  geom_bar(stat = "identity") +
  facet_grid(Region ~ Category) +
  labs(title = "Sales by Region, Category, and Subcategory",
       x = "Product Subcategory",
       y = "Total Sales") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 10, face = "bold"),
        strip.text = element_text(size = 10, face = "bold"))

Output:

facet_grid(Region ~ Category): Creates a matrix of plots split by both Region (rows) and Category (columns), allowing for a more comprehensive hierarchical view.

Implementing Hierarchical Visualization with Concatenation

For this example, we will use a hypothetical dataset representing sales data in different regions and sub-regions within countries. Each level of the hierarchy includes:

• Country (e.g., USA, Canada)
• Region (e.g., West, East)
• Sub-region (e.g., California, New York)

# Load necessary libraries
library(ggplot2)
library(dplyr)

# Create a sample dataset
data <- data.frame(
  Country = c("USA", "USA", "Canada", "Canada"),
  Region = c("West", "East", "West", "East"),
  SubRegion = c("California", "New York", "Alberta", "Ontario"),
  Sales = c(50000, 45000, 30000, 35000)
)
data

Output:

ountry	Region	SubRegion	Sales
<chr>	<chr>	<chr>	<dbl>
USA	West	California	50000
USA	East	New York	45000
Canada	West	Alberta	30000
Canada	East	Ontario	35000

In this example, Country, Region, and SubRegion represent the hierarchical structure we aim to display on the axis labels. Let’s begin by creating a simple bar plot of sales by sub-region without hierarchical axis labels.

# Basic bar plot
ggplot(data, aes(x = SubRegion, y = Sales)) +
  geom_bar(stat = "identity") +
  theme_minimal() +
  labs(title = "Sales by Sub-Region", x = "Sub-Region", y = "Sales")

Output:

This plot does not yet reflect the hierarchical relationship between countries, regions, and sub-regions.

Concatenating Hierarchical Labels

One of the simplest ways to represent hierarchical structures in ggplot2 is by concatenating the levels of hierarchy and displaying them as combined labels on the x-axis.In this code:

• We create a new variable HierarchicalLabel that concatenates Country, Region, and SubRegion into a single label using the paste function.
• This creates labels like USA > West > California, which clearly indicate the hierarchy. This method works well for smaller datasets or when the hierarchy is simple.

# Create a hierarchical label by concatenating levels
data <- data %>%
  mutate(HierarchicalLabel = paste(Country, Region, SubRegion, sep = " > "))

# Plot with hierarchical labels
ggplot(data, aes(x = HierarchicalLabel, y = Sales)) +
  geom_bar(stat = "identity") +
  theme_minimal() +
  labs(title = "Sales by Country, Region, and Sub-Region", x = "Location", y = "Sales") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

Output:

Displaying Hierarchy Using Faceting

For larger datasets, concatenating the labels may make the plot cluttered and hard to read. An alternative approach is to use faceting in ggplot2, which splits the data into multiple plots based on a categorical variable. We can facet the plot by higher levels of the hierarchy, like Country and Region, to separate the sub-regions into individual panels.

# Facet plot by Country and Region
ggplot(data, aes(x = SubRegion, y = Sales)) +
  geom_bar(stat = "identity") +
  theme_minimal() +
  labs(title = "Sales by Country, Region, and Sub-Region", x = "Sub-Region", y = "Sales") +
  facet_grid(Country ~ Region) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

Output:

In this example, facet_grid(Country ~ Region) splits the plot into panels based on Country and Region. Each panel contains the sub-regions specific to that combination. This method keeps the hierarchical structure clear without cluttering the x-axis.

Conclusion

Using ggplot2 to display hierarchical structures on axis labels can greatly enhance the readability and interpretability of your plots. By combining multiple levels of data into a single visualization, you can convey complex relationships and patterns effectively. With functions like facet_wrap() and facet_grid(), along with customized axis labels, you can create sophisticated, informative plots tailored to hierarchical datasets.