Visualizing Data with Seaborn - Mumbai West House Pricing

Customizing Seaborn Charts

Learning Outcome

Combine multiple plots to reveal both summary statistics and raw data

Enhance box plots to better interpret spread and outliers

Customize histograms for deeper distribution analysis

Create and enhance horizontal bar charts

Customize pie charts using Matplotlib alongside Seaborn

Explain why customization is necessary beyond default Seaborn plots

Previously covered concepts include:

Seaborn basics and core plot types

Using Seaborn with pandas DataFrames

Interpreting trends, distributions, and relationships

Basic plot customization (labels, hue, palette)

This topic extends those ideas by focusing on advanced visual refinement and combined visual techniques

Hook/Story/Analogy(Slide 4)

Transition from Analogy to Technical Concept(Slide 5)

Why Customization is Needed in Seaborn?

Definition

Customization refers to modifying visual properties such as colors, orientation, annotations, and overlays to improve interpretability and focus

Why this step matters?

Creating & Customizing Pie Charts

Note: Seaborn does not provide a dedicated pie chart function.
For proportion-based visuals, Matplotlib is used alongside Seaborn

Code to create:

import matplotlib.pyplot as plt

labels = ["Apartment", "Villa", "Studio"]
sizes = [60, 25, 15]

plt.pie(sizes, labels=labels, autopct="%1.1f%%")
plt.title("House Type Distribution")
plt.show()

Explanation:

sizes defines proportional values

autopct displays percentage contribution

Pie chart communicates part-to-whole relationship

Customizing Pie Charts:

explode = [0.1, 0, 0]
colors = ["skyblue", "lightgreen", "orange"]

plt.pie(
    sizes,
    labels=labels,
    autopct="%1.1f%%",
    explode=explode,
    colors=colors,
    startangle=140
)
plt.title("Customized House Type Distribution")
plt.show()

Customization Options:

explode – highlight a specific slice

colors – apply a consistent color scheme

startangle – rotate chart for better orientation

Creating and Customizing Horizontal Bar Charts

import seaborn as sns
import pandas as pd

data = pd.DataFrame({
    "Category": ["Apartment", "Villa", "Studio"],
    "Values": [120, 300, 80]
})

sns.barplot(x="Values", y="Category", data=data, orient="h")
plt.show()

Code to create:

Explanation:

Orientation is set explicitly to horizontal

Categories are displayed along the y-axis

Customizing Horizontal Bar Charts:

sns.barplot(
    x="Values",
    y="Category",
    data=data,
    orient="h",
    palette="coolwarm"
)

for i in range(len(data["Values"])):
    plt.text(data["Values"][i], i, data["Values"][i])

plt.xlim(0, 350)
plt.show()

Customization Options:

palette – control color theme

Axis limits using plt.xlim()

Annotating bars with exact values

Creating & Customizing Advanced Histograms

data = pd.DataFrame({
    "Values": [10, 20, 20, 30, 30, 30, 40, 40, 40, 40]
})

sns.histplot(data["Values"], bins=5)
plt.show()

Code to create:

Explanation:

Create a DataFrame with numeric values.

Plot their distribution using a 5-bin histogram.

Display the chart.

Histograms help understand how values are distributed across ranges

Customized Histogram with KDE

sns.histplot(data["Values"], bins=5, kde=True)
plt.show()

Customization Options:

bins – control granularity

kde – overlay density curve

hue – compare distributions

Why KDE?

Shows smooth distribution trend

Helps identify skewness and clustering

Creating and Customizing Advanced Box Plots

data = pd.DataFrame({
    "Category": ["A", "A", "B", "B", "C", "C"],
    "Values": [10, 20, 20, 30, 30, 40]
})

sns.boxplot(x="Category", y="Values", data=data)
plt.show()

Code to create:

Explanation:

Create a DataFrame with categories and numeric values.

Use sns.boxplot() to compare value distribution across categories.

Display the box plot.

Why box plots:

Summarize spread

Highlight medians and outliers

Customizing Box Plots:

sns.boxplot(
    x="Category",
    y="Values",
    data=data,
    notch=True,
    palette="Set2"
)
plt.show()

Customization Options:

notch=True – show confidence interval around median

palette – color differentiation

showfliers=False – hide outliers

Combining Plots for Enhanced Insights

sns.boxplot(x="Category", y="Values", data=data)
sns.swarmplot(x="Category", y="Values",
data=data, color="black")
plt.show()

Code :

Explanation:

Box plot provides statistical summary

Swarm plot shows individual data points

Together, they reveal both structure and detail

Why combine plots?

Box plots show summary

Swarm/strip plots show individual observations

Combined view improves interpretation

Summary

Combined plots provide richer analytical perspectives

Advanced histograms and box plots reveal deeper insights

Matplotlib complements Seaborn when required

Customization enhances clarity and focus

Quiz

Why are horizontal bar charts preferred for long category names?

A. Faster plotting

B. Better readability

C. Lower memory usage

D. Automatic sorting

Quiz-Answer

Why are horizontal bar charts preferred for long category names?

A. Faster plotting

B. Better readability

C. Lower memory usage

D. Automatic sorting