❮ Previous

Pair Plot & PairGrid

  • This module teaches how to create pair plots and PairGrids in Seaborn for detailed pairwise relationship analysis. You will learn multi-variable comparison, using hue, diagonal customization, custom pairwise plotting, and advanced mapping techniques for deeper insights in Python.

  • What is Pair Plot?

    Theory

    A Pair Plot:

    ✔ Shows pairwise relationships between numerical variables
    ✔ Displays scatter plots for variable combinations
    ✔ Shows distribution on diagonal
    ✔ Useful for correlation detection

    It is created using:

    sns.pairplot()


    Pairwise Relationship Analysis

    Example

Pairwise Relationship Analysis (Pair Plot)

This visualization uses Seaborn’s pairplot to show pairwise relationships between all numerical variables in the dataset. 

import seaborn as sns
import matplotlib.pyplot as plt

tips = sns.load_dataset("tips")

sns.pairplot(tips)
plt.show()
Lesson image

  • Output Explanation

    If dataset has 4 numeric columns:

    You get a 4x4 grid:

    • Diagonal → Distribution plots

    • Off-diagonal → Scatter plots

    • Shows relationship between every pair of variables

    You can analyze:

    ✔ Positive correlation
    ✔ Negative correlation
    ✔ No relationship
    ✔ Clusters


    Multi-variable Comparison

    Pairplot automatically compares:

    • total_bill vs tip

    • total_bill vs size

    • tip vs size

    • etc.

    This gives complete overview of numerical data.


    Using Hue

    Why Hue?

    To visualize category differences.

    Example

Pair Plot with Hue (Category Differences by Gender)

This visualization uses Seaborn’s pairplot with the hue parameter to explore pairwise relationships while distinguishing categories. 

sns.pairplot(tips, hue="sex")
plt.show()
Lesson image

  • Output Explanation

    • Different colors for Male & Female

    • See if relationship differs by gender

    • Observe clustering patterns

    You can also use:

sns.pairplot(tips, hue="smoker")

  • Diagonal Options

    By default, diagonal shows histogram.

    You can change it.

    Histogram (Default)

sns.pairplot(tips, diag_kind="hist")

  • KDE on Diagonal

sns.pairplot(tips, diag_kind="kde")

  • KDE gives smoother distribution view.


    Custom Pairwise Plotting (kind Parameter)

    You can control scatter type.

    Scatter (Default)

sns.pairplot(tips, kind="scatter")

  • Regression Line

sns.pairplot(tips, kind="reg")
Lesson image

  • Adds regression line in scatter plots.

    Useful for:

    • Trend analysis

    • Correlation strength


    What is PairGrid?

    Theory

    PairGrid is advanced version of pairplot.

    PairPlot

    PairGrid

    Simple & automatic

    Customizable

    Quick analysis

    Advanced control

    Less flexible

    Fully customizable

    Pairplot internally uses PairGrid.


    Advanced Custom Mapping (PairGrid)

    Basic Structure

Advanced Pairwise Analysis using PairGrid

This visualization uses Seaborn’s PairGrid for customized pairwise plotting. 

g = sns.PairGrid(tips)
g.map(sns.scatterplot)
plt.show()
Lesson image

  • Different Plots for Upper, Lower, Diagonal

    This is where PairGrid becomes powerful.

    Example

Custom PairGrid with Different Plots on Upper, Lower, and Diagonal

This visualization demonstrates the power of Seaborn’s PairGrid by using different plots for different parts of the grid. 

g = sns.PairGrid(tips)

g.map_upper(sns.scatterplot)
g.map_lower(sns.kdeplot)
g.map_diag(sns.histplot)

plt.show()
Lesson image

  • Output Explanation

    • Upper triangle → Scatter

    • Lower triangle → KDE

    • Diagonal → Histogram

    This creates a fully customized pairwise visualization.


    Using Hue in PairGrid

Custom PairGrid with Hue (Gender Comparison)

This visualization uses Seaborn’s PairGrid with the hue parameter to incorporate categorical differences into advanced pairwise analysis. 

g = sns.PairGrid(tips, hue="sex")

g.map_upper(sns.scatterplot)
g.map_lower(sns.kdeplot)
g.map_diag(sns.histplot)

g.add_legend()
plt.show()
Lesson image

  • Now categories are color-coded.

❮ Previous