Seaborn

0.13.2 · active · verified Sat Mar 28

Seaborn is a high-level Python library for creating statistical graphics, building on Matplotlib and integrating closely with Pandas data structures. It provides a dataset-oriented API to draw attractive and informative statistical plots with ease. The library is actively maintained with regular minor and major releases, currently at version 0.13.2, ensuring compatibility with evolving data science ecosystems.

Warnings

breaking The `seaborn.objects` interface was introduced in v0.12.0, offering a new declarative API. While powerful, it's considered experimental and may have breaking changes or rough edges in early minor releases.
Fix: Consult the official `seaborn.objects` documentation for specific usage and known limitations. For simpler plots or existing code, continue using the traditional axes-level and figure-level functions.
breaking Seaborn's categorical plotting functions (e.g., `boxplot`, `barplot`, `catplot`) underwent a major overhaul in v0.13.0. This includes changes to color defaults (now requiring explicit `hue` for multiple colors) and the introduction of the `native_scale` parameter.
Fix: Review code using categorical plots. To reproduce previous color behavior, explicitly assign a redundant `hue` variable (e.g., `boxplot(data, x='x', y='y', hue='x')`). If mixing categorical and numerical data on an axis, consider `native_scale=True`.
deprecated Python 3.7 support was dropped in Seaborn v0.12.2.
Fix: Upgrade to Python 3.8 or newer to ensure compatibility and receive future updates.
gotcha Positional arguments for most plotting functions were deprecated in v0.11.0 and enforced to be keyword-only in v0.12.0. This applies to arguments like `x`, `y`, `hue`, etc.
Fix: Always pass arguments like `x`, `y`, `data` using explicit keywords (e.g., `sns.scatterplot(data=df, x='col1', y='col2')` instead of `sns.scatterplot(df, 'col1', 'col2')`).
gotcha A regression in v0.13.0 caused exceptions when working with non-NumPy data types (e.g., Pandas nullable dtypes), which was fixed in v0.13.1. Pandas dtypes can still sometimes cause issues as Matplotlib (and by extension, Seaborn) often expects NumPy dtypes.
Fix: Update to Seaborn v0.13.1 or later. Ensure that data passed to Seaborn plotting functions are of appropriate (usually numerical) NumPy types, converting Pandas nullable dtypes explicitly if necessary (e.g., `.astype(float)`).
gotcha When combining categorical plots with other plot types (e.g., a bar plot and a line plot on the same axes), misalignment can occur because categorical plots internally map string/category values to integer indices.
Fix: For Seaborn v0.13.0+, use `native_scale=True` in categorical plots if the categorical axis truly represents numeric or datetime data. For older versions or string categories, carefully manage axis limits and tick labels using Matplotlib to ensure alignment.

Install

pip install seaborn Install stable version

Imports

seaborn
```
import seaborn as sns
```
The standard and widely adopted alias for Seaborn.
seaborn.objects
```
import seaborn.objects as so
```
Recommended alias for the experimental declarative 'objects' interface introduced in v0.12.0.

Quickstart

This quickstart demonstrates loading a built-in dataset, applying a theme, and creating a relational scatter plot using Seaborn's high-level API. It also includes basic Matplotlib calls for customization and displaying the plot.

import seaborn as sns
import matplotlib.pyplot as plt

sns.set_theme(style="darkgrid")

# Load an example dataset
tips = sns.load_dataset("tips")

# Create a scatter plot
sns.scatterplot(
    data=tips, 
    x="total_bill", 
    y="tip", 
    hue="smoker", 
    style="time", 
    size="size"
)

plt.title("Total Bill vs. Tip by Smoker and Time")
plt.xlabel("Total Bill ($)")
plt.ylabel("Tip ($)")
plt.show()

view raw JSON →