Create Faceted Grid of Seaborn Bar plots in Python

In this tutorial, you’ll learn how to use Seaborn to create faceted bar plots.

These plots allow you to break down complex data into sub-plots based on categorical variables.

We’ll start by understanding the basics of creating faceted grids and then move on to more advanced customization techniques.

 

 

Create Faceted Bar Plots

First, let’s create a sample dataset to work with:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
data = pd.DataFrame({
    'Region': ['North', 'South', 'East', 'West']*6,
    'Plan': ['Basic', 'Premium']*12,
    'Average_Usage': [450, 600, 550, 500, 700, 800, 750, 650, 480, 620, 580, 530, 720, 820, 770, 670, 460, 610, 560, 510, 710, 810, 760, 660]
})
print(data.head())

Output:

  Region    Plan  Average_Usage
0  North   Basic            450
1  South Premium            600
2   East   Basic            550
3   West Premium            500
...

Now, let’s create a faceted grid of bar plots using sns.FacetGrid. We will plot the average usage for each region, segmented by the type of plan (Basic or Premium).

g = sns.FacetGrid(data, col="Plan")
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
plt.show()

Output:

 

Adjust the Order of Facets

Let’s reorder the facets by the plan types.

g = sns.FacetGrid(data, col="Plan", col_order=["Premium", "Basic"])
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
plt.show()

Output:

This ordering places the ‘Premium’ plan before the ‘Basic’ plan.

 

Add Annotations and Labels

Let’s explore how to add informative annotations and labels to individual facets in our plots.

Add Annotations

Annotations can help highlight specific data points or trends in each facet. For instance, you might want to annotate the highest average usage in each plan type.

g = sns.FacetGrid(data, col="Plan", height=4, aspect=1.2)
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
for ax, title in zip(g.axes.flat, ["Premium", "Basic"]):
    # Find the maximum average usage in each facet
    max_usage = data[data['Plan'] == title]['Average_Usage'].max()
    max_region = data[(data['Average_Usage'] == max_usage) & (data['Plan'] == title)]['Region'].values[0]
    ax.text(x=max_region, y=max_usage, s=f'Max: {max_usage}', color='red', va='bottom')
plt.show()

Output:

The annotation ‘Max’ clearly marks the highest average usage in each plan type, making it easier for viewers to spot key data points.

Add Labels

Labels for each subplot can enhance the understandability of your faceted grid. Let’s add custom labels that provide insights about each plan.

g = sns.FacetGrid(data, col="Plan", height=4, aspect=1.2)
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
labels = {
    "Premium": "High Usage Plan",
    "Basic": "Economical Plan"
}
for ax, title in zip(g.axes.flat, ["Premium", "Basic"]):
    ax.set_title(labels[title])
plt.show()

Output:

 

Adjust the Scale and Aspect Ratio

Adjusting the scale and aspect ratio of individual facets in a Seaborn grid can significantly improve the readability and aesthetic appeal of your visualizations.

Adjust Aspect Ratio

You can adjust the aspect ratio of a plot:

g = sns.FacetGrid(data, col="Plan", height=4, aspect=1.5)
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
plt.show()

Output:

In this example, the elongated aspect ratio makes it easier to differentiate between the bars and discern the data.

Adjust Scale

You can also adjust the scale of the y-axis across facets for a consistent comparison, especially when dealing with varying ranges of data.

g = sns.FacetGrid(data, col="Plan", height=4, aspect=1.5, sharey=True)
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
plt.show()

Output:

With sharey=True, all facets now share the same y-axis scale, making it easier to compare the data across different plan types.

 

Using Different Seaborn Styles and Color Palettes

Applying different styles and color palettes can change the look and feel of your plots, making them more suitable for your specific context or audience.

Apply Different Seaborn Styles

Seaborn comes with several built-in themes that can be applied globally to all plots. Let’s try out a different style for our faceted grid.

sns.set_style("darkgrid")
g = sns.FacetGrid(data, col="Plan", height=4, aspect=1.5)
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
plt.show()

Output:

Using Color Palettes

Seaborn also allows for easy manipulation of color palettes, which can be very effective in differentiating elements within your plots.

sns.set_palette("viridis")
g = sns.FacetGrid(data, col="Plan", height=4, aspect=1.5)
g.map(sns.barplot, "Region", "Average_Usage", order=["North", "South", "East", "West"])
plt.show()

Output:

The ‘viridis’ palette offers a range of colors that are both visually appealing and distinct.