Data Visualization and Analysis With Seaborn Library/

...

/

Pair Plots

Pair Plots

Learn to design, customize, and interpret relationships among data features through pair plots.

We'll cover the following...

Overview

We use scatter plots for bivariate analysis of data. However, to see a correlation between more than two variables at once, there’s a more thoughtful way: pair plots. A pair plot displays pairwise relationships between the dataset’s variables and plots all numeric variables by default.

Use, customize and interpret pair plots

A pair plot is a great place to start when performing data analysis because it gives us a detailed initial view of our data.

We import the tips dataset from seaborn in the code snippet below. Let’s get started!

Press + to interact
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
sns.set_theme()    # set default theme
tips_df = sns.load_dataset('tips')
print(tips_df.head())

To construct a pair plot, we can use the sns.pairplot() function and pass our tips_df DataFrame to the data argument. We also pass True to the second argument, dropna, which means if our dataset has any null records, those will not be used in the visualization. How cool is that! The resulting pair plot is shown below:

Press + to interact
sns.pairplot(data = tips_df, dropna = True)

plt.savefig('output/graph.png')  # save figure

Isn’t it amazing how just one line of code generates such a powerful data visualization? We have three numeric variables in the tips dataset, creating nine pairwise plots. Let’s break down the pair plot visualization to understand the complete plot. First, as shown in ...