How To Align Kde Plot With Strip Plot In Seaborn?

A high-level interface for creating attractive and informative statistical graphics is offered by a powerful python library Seaborn. One of the most common tasks in data visualization is aligning different types of plots in one graph to gain insights into the data.

In this article, we will understand and implement how to align a KDE (Kernel Density Estimate) plot with a strip plot in seaborn.

Understanding KDE Plots and Strip Plots

Before aligning these plots is undertaken, let's briefly understand what they are and how they are used.

• KDE Plots: The underlying distribution of a dataset is visualized by a KDE plot, also known as a kernel density estimate plot, which is a statistical method. The data is smoothed, and patterns and trends that might be obscured by a histogram are identified.
• Strip Plots: The distribution of a single variable across different categories is displayed by a strip plot, which is a type of categorical plot. The relationship between a categorical variable and a continuous variable is visualized.

Aligning KDE and Strip Plot: Step-by-Step Examples

To align a KDE plot with a strip plot in seaborn, the sns.kdeplot() function is used for the KDE plot and the sns.stripplot() function is used for the strip plot.

Example 1: Overlapping KDE with Strip Plot on the Same Axes

In this basic approach:

• We use Seaborn's stripplot to show individual sepal lengths per species, with jitter for clarity.
• Then, we loop through each species and overlay KDE plots for each distribution.

This code visualizes the `sepal_length` distribution for different `species` in the iris dataset using a combined strip plot and KDE plot. The strip plot shows individual data points with some jitter for clarity, while the KDE plot overlays a smoothed density estimate for each species. This combination provides both detailed individual data points and an overall distribution view.

import seaborn as sns
import matplotlib.pyplot as plt
data = sns.load_dataset("iris")

fig, ax = plt.subplots(figsize=(8, 6))
# Plot a strip plot
sns.stripplot(x="species", y="sepal_length", data=data, ax=ax, jitter=True, color="blue", alpha=0.6)
# Plot a KDE plot for each species
species = data['species'].unique()
for sp in species:
    subset = data[data['species'] == sp]
    sns.kdeplot(subset['sepal_length'], ax=ax, label=sp, fill=True, alpha=0.4)
plt.show()

Output:

Example 2: Strip Plot and KDE Plot with Shared Y-Axis

This code visualizes the total_bill distribution across different days in the tips dataset using a combined strip plot and KDE plot.

• The strip plot displays individual data points with some jitter for clarity on ax1,
• while the KDE plot overlays smoothed density estimates for each day on ax2, which shares the same y-axis.

This combination allows for a detailed view of individual data points and their overall density distribution on the same plot.

import seaborn as sns
import matplotlib.pyplot as plt
data = sns.load_dataset("tips")
fig, ax1 = plt.subplots(figsize=(8, 6))

# Plot a strip plot
sns.stripplot(x="day", y="total_bill", data=data, ax=ax1, jitter=True, color="blue", alpha=0.6)
# Create a twin axis sharing the same y-axis
ax2 = ax1.twinx()
# Plot a KDE plot on the twin axis
days = data['day'].unique()
for day in days:
    subset = data[data['day'] == day]
    sns.kdeplot(subset['total_bill'], ax=ax2, label=day, fill=True, alpha=0.4)
plt.show()

Output:

Example 3: Customizing Colors and Styles in Kde Plot With Strip Plot

In this example, we will customize the colors and styles of the KDE and strip plots to make the visualization more appealing and informative.

• Strip Plot: Uses the palette parameter to apply a custom color palette and edgecolor and linewidth to outline the data points.
• KDE Plot: Customizes the line style and width using linestyle and linewidth.
• Customization: Adds a title, labels, and a legend for better readability.

import seaborn as sns
import matplotlib.pyplot as plt
iris = sns.load_dataset("iris")

# Create a strip plot with custom colors
sns.stripplot(x="species", y="sepal_length", data=iris, jitter=True, palette="Set2", edgecolor="black", linewidth=1)

# Overlay KDE plots with custom styles
for species in iris['species'].unique():
    subset = iris[iris['species'] == species]
    sns.kdeplot(subset['sepal_length'], label=species, bw_adjust=0.5, linestyle="--", linewidth=2)

plt.title("Customized Sepal Length Distribution by Species")
plt.xlabel("Species")
plt.ylabel("Sepal Length")
plt.legend(title='Species')
plt.show()

Output:

Conclusion

Seaborn provides a powerful interface for creating visually appealing and informative statistical graphics. By combining KDE (Kernel Density Estimate) plots with strip plots, we can gain deeper insights into our data by visualizing both the underlying distributions and the relationship between categorical and continuous variables.