Search⌘ K
AI Features
Home
Courses
Data Visualization and Analysis With Seaborn Library

Visualization with Heatmaps

Explore how to use Seaborn to create heatmaps that visualize correlations and structured numerical data. Learn techniques to handle multilevel indexes, style color palettes, customize labels, and annotate heatmaps for clearer data interpretation.

We'll cover the following...

Overview

A heatmap is a visual approach to displaying a data table. It represents the rectangular data as a color-encoded matrix. The color maps use color variation to represent different details by varying hue, saturation, or brightness.

Correlation with heatmaps

We begin by importing the required libraries (seaborn, pandas, matplotlib) and set the default seaborn theme using the sns.set_theme() function. Next, we import the penguins dataset and view the data with the pandas head() function, as shown below:

Python
import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt
sns.set_theme()
penguins_df = sns.load_dataset('penguins')
print(penguins_df.head())

Heatmaps are popular for representing correlation among variables. We plot a correlation matrix of the penguins dataset on a heatmap by passing penguins_df.corr() to the sns.heatmap() function. We’ve customized the font size for the plot by specifying sns.set(font_scale=0.7) so that the complete column names are visible in the plot. The heatmap is shown below:

Python
sns.set(font_scale=0.7)
sns.heatmap(penguins_df.corr())
plt.savefig('output/graph.png')

The corr() function computes Pearson’s correlation for numeric variables present in the data. The values range from –1 to +1. The negative values represent a negative correlation, and the positive values represent a positive correlation.

In a heatmap, as shown through the color bar, low correlation values are ...