seaborn

seaborn builds statistical graphics on matplotlib with tidy-data APIs - ideal for fast exploratory analysis and polished defaults.

Recipe

Quick-reference recipe card - copy-paste ready.

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
 
sns.set_theme(style="whitegrid", context="notebook")
df = pd.read_csv("sales.csv", parse_dates=["ordered_at"])
g = sns.relplot(data=df, x="ordered_at", y="revenue", hue="region", kind="line")
g.set_axis_labels("Date", "Revenue ($)")
plt.close()

When to reach for this:

• Quick EDA on pandas DataFrames
• Group comparisons with confidence intervals
• Faceted small multiples (col, row)
• KDE/hist/box/violin for distributions

Working Example

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
 
rng = np.random.default_rng(42)
df = pd.DataFrame(
    {
        "region": np.repeat(["East", "West"], 200),
        "segment": np.tile(np.repeat(["SMB", "Enterprise"], 100), 2),
        "revenue": rng.normal(150, 40, 400).clip(10),
    }
)
 
sns.set_theme(style="ticks")
fig, axes = plt.subplots(1, 3, figsize=(12, 4))
 
sns.histplot(data=df, x="revenue", hue="region", kde=True, ax=axes[0])
axes[0].set_title("Distribution")
 
sns.boxplot(data=df, x="segment", y="revenue", hue="region", ax=axes[1])
axes[1].set_title("By segment")
 
sns.barplot(
    data=df,
    x="region",
    y="revenue",
    estimator="median",
    errorbar=("pi", 50),
    ax=axes[2],
)
axes[2].set_title("Median with interval")
 
sns.despine()
fig.tight_layout()
fig.savefig("eda.png", dpi=150, bbox_inches="tight")
plt.close(fig)

What this demonstrates:

• Tidy data= API across three plot types
• histplot with optional KDE overlay
• boxplot with hue for second categorical
• Modern errorbar tuple instead of deprecated ci

Deep Dive

How It Works

• seaborn maps DataFrame columns to aesthetic roles (x, y, hue, size, style).
• Statistical functions aggregate internally (mean, median, bootstrap CI).
• Figure-level functions (relplot, catplot, displot) return FacetGrid.
• Axes-level functions (scatterplot, barplot) accept ax= for subplots.

Plot Picker

Python Notes

import seaborn as sns
 
# Load sample dataset for teaching
tips = sns.load_dataset("tips")  # bundled CSV cache
 
# Palette accessible to colorblind readers
sns.set_palette("colorblind")

Gotchas

• Wide vs tidy data - seaborn expects one row per observation. Fix: melt or fix upstream schema.
• Deprecated ci= parameter - removed in seaborn 0.13+. Fix: use errorbar=("ci", 95) or None.
• Forgotten plt.close on relplot - figure-level APIs create new figures silently. Fix: capture return value and plt.close().
• Bar plot of means on skewed data - mean hides outliers. Fix: pair with box/violin or show median.
• Huge categorical hue - unreadable legends. Fix: top-N filter or facet instead of hue.

Alternatives

FAQs

sns.relplot(data=df, x="x", y="y", col="region", row="segment", kind="scatter")

sns.lineplot(data=df, x="t", y="v", errorbar=None)

g = sns.relplot(...)
g.savefig("out.png", dpi=150)

sns.regplot(data=df, x="ad_spend", y="revenue")

order = df.groupby("region")["revenue"].median().sort_values().index
sns.barplot(data=df, x="region", y="revenue", order=order)

Related

• matplotlib - axes-level customization
• Visualization Basics - chart selection
• Plotly - interactive follow-ups
• Visualization Best Practices - honest scales

Stack versions: This page was written for Python 3.14.0 (stable 3.14, maintenance 3.13), FastAPI 0.115+, Django 5.2, Flask 3.1, Pydantic 2, PyTorch 2.6+, pandas 2.2+, Polars 1.x, ruff 0.9+, and uv 0.6+.