Pair plot

seaborn.pairplot() creates a grid of scatterplots and histograms (or KDEs) for pairwise relationships between numeric variables in a dataset.

It’s one of Seaborn’s most powerful exploratory tools, showing how each variable:

• correlates with others (scatterplots), and
• distributes individually (histograms or density plots on the diagonal).

Syntax:

sns.pairplot(    data,    *,    hue=None,    vars=None,    x_vars=None,    y_vars=None,    kind="scatter",    diag_kind="auto",    markers=None,    palette=None,    corner=False,    plot_kws=None,    diag_kws=None,    grid_kws=None,    height=2.5,    aspect=1,    dropna=True,    context=None,)

Parameters:

• data = DataFrame with numeric columns
• hue = Variable to colour-code data points by category
• vars = List of variables to include (both x and y)
• x_vars, y_vars = Lists of variables for x and y axes (for asymmetric grids)
• kind = Plot type: "scatter" (default) or "kde"
• diag_kind = Plot type on the diagonal: "auto", "hist", or "kde"
• markers = Marker style for scatterplots
• palette = Colour palette for hue categories
• corner = If True, show only the lower triangle of plots
• plot_kws, diag_kws, grid_kws = Keyword arguments for customising subplots
• height = Height (in inches) of each subplot
• aspect = Aspect ratio of each subplot
• dropna = Whether to drop missing values
• context = Set a plotting context (e.g., "talk", "notebook")

Basic example

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris)plt.show()

Creates a scatterplot matrix for all numeric columns in iris. Diagonal = histograms of each variable. Off-diagonal = pairwise scatterplots.

Add hue (colour by category)

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, hue="species", palette="Set2")plt.show()

Colours each species differently. Great for class separation or cluster analysis.

Use KDE instead of scatterplots

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, hue="species", kind="kde")plt.show()

Shows smooth density contours instead of discrete points. Great for large datasets or overlapping points.

Change diagonal plot type

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, hue="species", diag_kind="kde")plt.show()

Replaces histograms on the diagonal with KDE (smoothed density) curves.

Select specific variables

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, vars=["sepal_length", "sepal_width", "petal_length"], hue="species")plt.show()

Only plots relationships among selected columns. Useful for focusing on variables of interest.

Asymmetric grids (x_vars vs y_vars)

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(    iris,    x_vars=["sepal_length", "sepal_width", "petal_length"],    y_vars=["sepal_width", "petal_width"],    hue="species")plt.show()

Creates a non-square grid — useful for comparing two sets of variables.

Show only lower triangle (corner)

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, hue="species", corner=True)plt.show()

Hides the upper triangle, reducing redundancy. Common in publications for clarity.

Change marker style

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, hue="species", markers=["o", "s", "D"])plt.show()

Uses different marker shapes for each category in hue.

Pass custom plotting options

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(    iris,    hue="species",    plot_kws={"alpha": 0.7, "s": 60, "edgecolor": "k"},    diag_kws={"fill": True, "linewidth": 2})plt.show()

Adds transparency, size, and edges to scatter points. Smooths and thickens KDEs/histograms on the diagonal.

Pairplot with regression lines

import seaborn as snsimport matplotlib.pyplot as plt​iris = sns.load_dataset("iris")​sns.pairplot(iris, kind="reg", hue="species")plt.show()

Adds linear regression lines in each scatterplot. Equivalent to calling sns.lmplot() for every pair.