Seaborn Cheat Sheet

Complete guide for statistical data visualization with Seaborn, the Python data visualization library that provides a high-level interface for creating attractive statistical graphics.

Core Seaborn Functions

Action	Code Example	Description
Import and Set Style	`import seaborn as sns` `sns.set(style="darkgrid")`	Imports the library and sets a default aesthetic style for all subsequent plots, including those made with pandas or Matplotlib.
Histogram with KDE	`sns.histplot(data, bins=50, kde=True)`	Plots a histogram of the data distribution. The `kde=True` argument overlays a kernel-density estimate.
2D Kernel-Density Plot	`sns.kdeplot(x=data1, y=data2)`	Creates a contour plot showing the joint kernel-density estimate for two variables.
Joint Distribution Plot	`sns.jointplot(x=data1, y=data2, kind="hex")`	Visualizes the joint distribution of two variables along with their individual marginal distributions. `kind="hex"` creates a hexbin plot.
Box Plot	`sns.boxplot(my_dataframe, palette="pastel")`	Creates a standard boxplot showing descriptive statistics (median, quartiles, outliers) for one or more variables.
Violin Plot	`sns.violinplot(my_dataframe, palette="pastel")`	Combines a boxplot with a kernel-density estimate on each side, showing the distribution of the data.
Categorical Violin Plot	`sns.violinplot(x=df.index.month, y=df.outdoor)`	Groups the data by a categorical variable (`x`) and creates a violin plot for the numerical variable (`y`) for each category.
Heatmap	`sns.heatmap(pivot_table, ax=ax)`	Visualizes a 2D matrix of data (like a pivot table) as a color-encoded grid.

Advanced Plot Types

Statistical Plots

# Scatter plots with linear regression
sns.regplot(x="column1", y="column2", data=df,
            line_kws={"color": "red"}, scatter_kws={"alpha": 0.5})

# Plot data and confidence intervals
sns.lineplot(data=df, x="time", y="value",
             hue="category", style="method")

# Point plots with confidence intervals
sns.pointplot(data=df, x="parameter", y="score",
              hue="dataset", dodge=True)

Categorical Plots

# Bar plots for categorical data
sns.barplot(data=df, x="category", y="value", hue="group")

# Strip plots (scatter plot for categorical data)
sns.stripplot(data=df, x="category", y="value", jitter=True)

# Swarm plots (non-overlapping points)
sns.swarmplot(data=df, x="category", y="value", hue="group")

# Count plots (histogram for categorical data)
sns.countplot(data=df, x="category", hue="group")

Pairwise Relationships

# Pair plot for multiple variable relationships
sns.pairplot(df, hue="species", vars=["sepal_length", "sepal_width"])

# Pair grid with custom plots
g = sns.PairGrid(df, hue="species")
g.map_upper(sns.scatterplot)
g.map_lower(sns.kdeplot)
g.map_diag(sns.histplot)
g.add_legend()

Styling and Customization

Color Palettes

# Built-in color palettes
sns.color_palette("hls", 8)
sns.color_palette("viridis", as_cmap=True)

# Custom palette from hex colors
colors = ["#FC8EAC", "#F76C6C", "#A8D8B9"]
sns.set_palette(sns.color_palette(colors))

# Diverging palettes
sns.diverging_palette(220, 20, n=10)

Figure Styles

# Set overall style
sns.set_style("whitegrid")
sns.set_style("darkgrid")
sns.set_style("ticks")

# Context for different presentations
sns.set_context("paper")    # Small plots for publication
sns.set_context("notebook") # Standard notebook plots
sns.set_context("talk")     # Larger plots for presentations

Themes and Font Settings

# Custom theme dictionary
custom_theme = {
    "font.style": "normal",
    "text.color": "green",
    "axes.labelcolor": "green",
    "axes.edgecolor": "green"
}
sns.set_theme(style="white", rc=custom_theme)

# Font scale
sns.set_theme(font_scale=1.5)

Working with pandas DataFrames

Integration with pandas

# Load and explore data
import pandas as pd
tips = sns.load_dataset("tips")

# Get first few rows
print(tips.head())

# Check data types
print(tips.dtypes)

# Summary statistics
print(tips.describe())

Exploratory Analysis Flow

# 1. Overview of variables
sns.histplot(data=tips, x="total_bill", kde=True)

# 2. Bivariate relationships
sns.scatterplot(data=tips, x="total_bill", y="tip", hue="smoker")

# 3. Categorical vs Numerical
sns.boxplot(data=tips, x="day", y="total_bill")
sns.violinplot(data=tips, x="day", y="total_bill")

# 4. Correlations
numerical_cols = ["total_bill", "tip", "size"]
corr = tips[numerical_cols].corr()
sns.heatmap(corr, annot=True, cmap="coolwarm")

# 5. Pairwise relationships
numerical_cols = tips.select_dtypes(include='number').columns
sns.pairplot(tips[numerical_cols])

Specialized Plots

Time Series Visualization

# Line plots with confidence intervals
sns.lineplot(data=stocks, x="date", y="price", hue="company")

# Seasonal decomposition plot
from statsmodels.tsa.seasonal import seasonal_decompose
result = seasonal_decompose(data, model='additive', period=365)
fig, (ax1, ax2, ax3, ax4) = plt.subplots(4, 1, figsize=(12, 10))
result.observed.plot(ax=ax1)
result.trend.plot(ax=ax2)
result.seasonal.plot(ax=ax3)
result.resid.plot(ax=ax4)

Statistical Plotting

# Residual plots
from scipy import stats
fig, ax = plt.subplots(figsize=(8, 6))
sns.residplot(x="x", y="y", data=dataframe, ax=ax)
ax.set_title("Residuals Plot")

# QQ plots
stats.probplot(data, dist="norm", plot=plt)
plt.title("Q-Q Plot")

# Correlation matrix
correlation_matrix = dataframe.corr()
sns.heatmap(correlation_matrix, annot=True, cmap="RdYlGn")

Faceted Plots

# FacetGrid for subset plotting
g = sns.FacetGrid(tips, col="time", row="smoker")
g.map(sns.scatterplot, "total_bill", "tip")

# JointGrid with marginal plots
g = sns.JointGrid(data=tips, x="total_bill", y="tip")
g.plot(sns.scatterplot, sns.histplot)

# PairGrid for multiple relationships
g = sns.PairGrid(tips, vars=["total_bill", "tip", "size"])
g.map_upper(sns.scatterplot)
g.map_lower(sns.kdeplot)
g.map_diag(sns.histplot)

Essential Seaborn Workflow

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

# 1. Set up the aesthetics
sns.set_theme(style="whitegrid", context="talk")

# 2. Load or prepare data
df = sns.load_dataset("tips")

# 3. Create plots
fig, axes = plt.subplots(2, 2, figsize=(12, 10))

# Distribution plots
sns.histplot(df["total_bill"], ax=axes[0,0])
axes[0,0].set_title("Bill Distribution")

# Categorical plots
sns.boxplot(data=df, x="day", y="total_bill", ax=axes[0,1])
axes[0,1].set_title("Bills by Day")

# Scatter plots
sns.scatterplot(data=df, x="total_bill", y="tip", hue="time", ax=axes[1,0])
axes[1,0].set_title("Bills vs Tips")

# Correlation
df_numeric = df.select_dtypes(include=[np.number])
correlation_matrix = df_numeric.corr()
sns.heatmap(correlation_matrix, ax=axes[1,1], annot=True, cmap="coolwarm")
axes[1,1].set_title("Correlation Matrix")

plt.tight_layout()
plt.show()

Best Practices

Common Mistakes to Avoid

Don't overload plots with too many elements
Use appropriate color schemes for accessibility
Ensure clear labels and legends
Choose plot types that best represent your data
Handle missing data appropriately before plotting

Performance Tips

# For large datasets, use sampling or aggregation
df_sample = df.sample(n=1000)

# Use vectorized operations
df["calculated_column"] = df["col1"] * df["col2"]

# Avoid for loops in favor of vectorized operations
# Good: df["new_col"] = df["col"] * 2
# Bad: for i in range(len(df)): df.loc[i, "new_col"] = df.loc[i, "col"] * 2

Seaborn excels at creating beautiful statistical visualizations with minimal code. By understanding its core functions and customization capabilities, you can create publication-quality plots quickly and effectively.

Updated: January 15, 2025
Author: Danial Pahlavan
Category: Data Science & Visualization