Python Data Analysis Deep Dive — Complete Guide to Data Visualization with matplotlib, seaborn & plotly

Lecture 3 Overview

There’s a saying that “one chart beats a thousand tables.” Data visualization is a core skill for spotting patterns and communicating insights to stakeholders. We’ll build the fundamentals with matplotlib, handle statistical visualization with seaborn, and build interactive dashboards with plotly.

---

Which Chart Should You Use?

Choosing the right chart is the single most important decision in visualization.

Goal	Recommended chart	Library
See a distribution	Histogram, KDE, box plot	seaborn
Compare	Bar chart, box plot	seaborn, matplotlib
Explore a relationship	Scatter plot, heatmap	seaborn
Time series trend	Line chart	matplotlib, plotly
Proportion	Pie chart, stacked bar	matplotlib
Dashboard	Interactive chart	plotly

---

1. matplotlib Basics

import matplotlib.pyplot as plt
import numpy as np

# A basic line chart
fig, ax = plt.subplots(figsize=(10, 6))
ax.plot(x, y, color='steelblue', linewidth=2, label='Actuals')
ax.set_title('Monthly Revenue Trend', fontsize=14, fontweight='bold')
ax.set_xlabel('Month')
ax.set_ylabel('Revenue (in thousands)')
ax.legend()
ax.grid(True, alpha=0.3)
plt.tight_layout()
plt.show()

# Multiple subplots
fig, axes = plt.subplots(2, 2, figsize=(12, 8))
axes[0, 0].hist(df['age'], bins=30, color='steelblue', alpha=0.7)
axes[0, 0].set_title('Age Distribution')
axes[0, 1].scatter(df['age'], df['income'], alpha=0.5)
axes[0, 1].set_title('Age vs. Income')
plt.tight_layout()

---

2. seaborn: Statistical Visualization

seaborn runs on top of matplotlib and is optimized for statistical visualization.

import seaborn as sns

# Histogram + KDE
sns.histplot(df['score'], kde=True, bins=30)

# Box plot — distribution and outliers at a glance
sns.boxplot(x='dept', y='score', data=df, palette='Set2')

# Violin plot — also shows the shape of the distribution
sns.violinplot(x='dept', y='score', data=df)

# Scatter plot + regression line
sns.regplot(x='experience', y='salary', data=df, scatter_kws={'alpha': 0.5})

# Correlation heatmap — an EDA essential
corr = df.corr()
sns.heatmap(corr, annot=True, fmt='.2f', cmap='RdYlGn', center=0)

# pairplot — every pairwise relationship at once
sns.pairplot(df[['age', 'income', 'score', 'dept']], hue='dept')

FacetGrid — Multiple Plots by Condition

g = sns.FacetGrid(df, col='dept', hue='gender', height=4)
g.map(sns.scatterplot, 'experience', 'salary')
g.add_legend()

---

3. plotly: Interactive Visualization

Use plotly for dashboards or interactive exploration in a Jupyter notebook.

import plotly.express as px
import plotly.graph_objects as go

# A basic scatter plot
fig = px.scatter(df, x='experience', y='salary', color='dept',
                 hover_data=['name'], size='score',
                 title='Experience vs. Salary')
fig.show()

# A time series line chart
fig = px.line(df_time, x='date', y='value', color='category',
              title='Trend Over Time by Category')
fig.update_traces(mode='lines+markers')

# An interactive bar chart
fig = px.bar(df.groupby('dept')['score'].mean().reset_index(),
             x='dept', y='score', color='dept',
             title='Average Score by Department')
fig.show()

---

A Practical Visualization Checklist

1. Titles and labels: Add clear descriptions to every axis and legend.
2. Color: Use colorblind-friendly palettes (colorblind or Set2).
3. Scale: A y-axis that doesn’t start at zero can mislead readers.
4. Density vs. frequency: Use density when comparing proportions.
5. Resolution: plt.savefig('plot.png', dpi=300, bbox_inches='tight')

---

Frequently Asked Questions

Q: Should I learn seaborn or plotly? A: seaborn for static images meant for reports and papers; plotly for web dashboards and exploratory analysis. In real data analysis work, you’ll need both.

Q: Is visualizing large datasets (a million rows) slow? A: Plotting a million points is slow. Summarize first — sampling, aggregation, heatmaps, or 2D histograms — before visualizing. Pairing plotly with datashader is another option.