Summary  
This chapter covers generating a heatmap from a 2D dataset to visualize pairwise correlations and shows how to customize annotations, color maps, color bar display, and tick rotations for improved readability.

General domain of usage  
Exploratory data analysis

A **heatmap** is a method for visualizing two-dimensional data using colors to represent the magnitude of each value. 


Definition

This example uses a heatmap to visualize **pairwise correlations** between variables.

## Creating a Simple Heatmap

`seaborn.heatmap()` takes a 2D dataset. A common use case is plotting a **correlation matrix**: given a `DataFrame`, call `.corr()` to compute correlations, then pass the resulting matrix to `heatmap()`.

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd

url = 'https://content-media-cdn.codefinity.com/courses/47339f29-4722-4e72-a0d4-6112c70ff738/countries_data.csv'
countries_df = pd.read_csv(url, index_col=0)

correlation_matrix = countries_df.corr(numeric_only=True)

sns.heatmap(correlation_matrix)
plt.show()

The correlation matrix is created from **numeric columns** only (`numeric_only=True`).

## Annotation and Colors

Setting `annot=True` writes the correlation values inside each cell. We can also pick a colormap using `cmap`.

It is also possible to change the **colors** for our heatmap via setting the `cmap` parameter (you can explore it in the <a href="https://seaborn.pydata.org/tutorial/color_palettes.html" target="_blank"><svg width="1em" height="1em" viewBox="0 0 30 32" fill="none" xmlns="http://www.w3.org/2000/svg"><path fill="none" stroke="#098f67" style="stroke: var(--color1, #098f67)" stroke-linejoin="miter" stroke-linecap="round" stroke-miterlimit="4" stroke-width="2.4156" d="M17.289 17.305v0c-1.754 1.754-4.597 1.754-6.351 0l-7.76-7.759c-1.755-1.755-1.755-4.601 0-6.356v0c1.753-1.753 4.595-1.756 6.351-0.005l6.208 6.187"></path><path fill="none" stroke="#098f67" style="stroke: var(--color1, #098f67)" stroke-linejoin="miter" stroke-linecap="round" stroke-miterlimit="4" stroke-width="2.4156" d="M12.504 13.97v0c1.754-1.754 4.597-1.754 6.351 0l7.762 7.762c1.754 1.754 1.754 4.597 0 6.351v0c-1.754 1.754-4.597 1.754-6.351 0l-5.953-5.953"></path></svg> "Choosing color palettes"</a> article). 

Note

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd

url = 'https://content-media-cdn.codefinity.com/courses/47339f29-4722-4e72-a0d4-6112c70ff738/countries_data.csv'
countries_df = pd.read_csv(url, index_col=0)

correlation_matrix = countries_df.corr(numeric_only=True)

sns.heatmap(correlation_matrix, annot=True, cmap='viridis')
plt.show()

The **color bar** on the right can be removed by setting `cbar=False`.


In most of the cases that's all you will need from a **heatmap customization**, however, you can always explore more in <a href="https://seaborn.pydata.org/generated/seaborn.heatmap.html" target="_blank"><svg width="1em" height="1em" viewBox="0 0 30 32" fill="none" xmlns="http://www.w3.org/2000/svg"><path fill="none" stroke="#098f67" style="stroke: var(--color1, #098f67)" stroke-linejoin="miter" stroke-linecap="round" stroke-miterlimit="4" stroke-width="2.4156" d="M17.289 17.305v0c-1.754 1.754-4.597 1.754-6.351 0l-7.76-7.759c-1.755-1.755-1.755-4.601 0-6.356v0c1.753-1.753 4.595-1.756 6.351-0.005l6.208 6.187"></path><path fill="none" stroke="#098f67" style="stroke: var(--color1, #098f67)" stroke-linejoin="miter" stroke-linecap="round" stroke-miterlimit="4" stroke-width="2.4156" d="M12.504 13.97v0c1.754-1.754 4.597-1.754 6.351 0l7.762 7.762c1.754 1.754 1.754 4.597 0 6.351v0c-1.754 1.754-4.597 1.754-6.351 0l-5.953-5.953"></path></svg> <code>heatmap()</code> documentation</a>.

Study More

## Improving Readability

The final thing that would improve the **readability** of our heatmap is rotating the ticks using already familiar `xticks()` and `yticks()` functions:

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd

url = 'https://content-media-cdn.codefinity.com/courses/47339f29-4722-4e72-a0d4-6112c70ff738/countries_data.csv'
countries_df = pd.read_csv(url, index_col=0)

correlation_matrix = countries_df.corr(numeric_only=True)

sns.heatmap(correlation_matrix, annot=True, cmap='viridis')
plt.xticks(rotation=20)
plt.yticks(rotation=20)

plt.show()

import unittest
import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import user_code

def _dynamic_test(test_case, condition, success_message, failure_message):
    if condition:
        test_case._testMethodName = success_message
        test_case.assertTrue(True, success_message)
    else:
        test_case._testMethodName = failure_message
        test_case.fail(failure_message)

class TestCorrelationHeatmap(unittest.TestCase):
    def setUp(self):
        self.df = sns.load_dataset('penguins')
        self.expected_corr = self.df.corr(numeric_only=True)

    def test_corr_matrix_correct(self):
        actual_corr = getattr(user_code, 'correlation_matrix', None)
        _dynamic_test(
            self,
            isinstance(actual_corr, pd.DataFrame) and actual_corr.equals(self.expected_corr),
            "Correlation matrix created correctly",
            "Correlation matrix not created correctly or incorrect parameters"
        )

    def test_heatmap_used(self):
        ax = plt.gca()
        mesh = None
        for collection in ax.collections:
            if hasattr(collection, 'get_array'):
                mesh = collection
                break
        _dynamic_test(
            self,
            mesh is not None,
            "Heatmap function used correctly",
            "Heatmap function not detected"
        )

    def test_tick_rotation(self):
        xtick_rotations = [label.get_rotation() for label in plt.gca().get_xticklabels()]
        ytick_rotations = [label.get_rotation() for label in plt.gca().get_yticklabels()]
        _dynamic_test(
            self,
            all(rot == 15 for rot in xtick_rotations) and all(rot == 15 for rot in ytick_rotations),
            "Tick labels rotated correctly",
            "Tick label rotation is incorrect"
        )

if __name__ == '__main__':
    unittest.main()


test_heatmap.py

Learn how to explore and communicate data through visual analysis. The course covers creating common plots, customizing visualizations, and using statistical charts to examine data distributions, relationships, and patterns. Exploratory data analysis techniques are applied to identify trends, anomalies, and insights in real-world datasets.

Heatmap

Creating a Simple Heatmap

Annotation and Colors

Improving Readability

Solution