/data-visualization

name: Data Visualization
description: Create effective visualizations using matplotlib and seaborn for exploratory analysis, presenting insights, and communicating findings with business stakeholders

Data Visualization

Overview

Data visualization transforms complex data into clear, compelling visual representations that reveal patterns, trends, and insights for storytelling and decision-making.

When to Use

• Exploratory data analysis and pattern discovery
• Communicating insights to stakeholders
• Comparing distributions and relationships
• Presenting findings in reports and dashboards
• Identifying outliers and anomalies visually
• Creating publication-ready charts and graphs

Visualization Types

• **Distributions**: Histograms, KDE, violin plots
• **Relationships**: Scatter plots, line plots, heatmaps
• **Comparisons**: Bar charts, box plots, ridge plots
• **Compositions**: Pie charts, stacked bars, treemaps
• **Temporal**: Line plots, area charts, time series
• **Multivariate**: Pair plots, correlation heatmaps

Design Principles

• Choose appropriate chart type for data
• Minimize ink-to-data ratio
• Use color purposefully
• Label clearly and completely
• Maintain consistent scales
• Consider accessibility

Implementation with Python

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from matplotlib.gridspec import GridSpec

# Set style
sns.set_style("whitegrid")
plt.rcParams['figure.figsize'] = (12, 6)

# Generate sample data
np.random.seed(42)
n = 500
data = pd.DataFrame({
    'age': np.random.uniform(20, 70, n),
    'income': np.random.exponential(50000, n),
    'education_years': np.random.uniform(12, 20, n),
    'category': np.random.choice(['A', 'B', 'C'], n),
    'region': np.random.choice(['North', 'South', 'East', 'West'], n),
    'satisfaction': np.random.uniform(1, 5, n),
    'purchased': np.random.choice([0, 1], n),
})

print(data.head())

# 1. Distribution Plots
fig, axes = plt.subplots(2, 2, figsize=(12, 8))

# Histogram
axes[0, 0].hist(data['age'], bins=30, color='skyblue', edgecolor='black')
axes[0, 0].set_title('Age Distribution (Histogram)')
axes[0, 0].set_xlabel('Age')
axes[0, 0].set_ylabel('Frequency')

# KDE plot
data['income'].plot(kind='kde', ax=axes[0, 1], color='green', linewidth=2)
axes[0, 1].set_title('Income Distribution (KDE)')
axes[0, 1].set_xlabel('Income')

# Box plot
sns.boxplot(data=data, y='satisfaction', x='category', ax=axes[1, 0], palette='Set2')
axes[1, 0].set_title('Satisfaction by Category (Box Plot)')

# Violin plot
sns.violinplot(data=data, y='age', x='category', ax=axes[1, 1], palette='Set2')
axes[1, 1].set_title('Age by Category (Violin Plot)')

plt.tight_layout()
plt.show()

# 2. Relationship Plots
fig, axes = plt.subplots(2, 2, figsize=(12, 8))

# Scatter plot
axes[0, 0].scatter(data['age'], data['income'], alpha=0.5, s=30)
axes[0, 0].set_title('Age vs Income (Scatter Plot)')
axes[0, 0].set_xlabel('Age')
axes[0, 0].set_ylabel('Income')

# Scatter with regression line
sns.regplot(x='age', y='income', data=data, ax=axes[0, 1], scatter_kws={'alpha': 0.5})
axes[0, 1].set_title('Age vs Income (with Regression Line)')

# Joint plot alternative
ax_hex = axes[1, 0]
hexbin = ax_hex.hexbin(data['age'], data['income'], gridsize=15, cmap='YlOrRd')
ax_hex.set_title('Age vs Income (Hex Bin)')
ax_hex.set_xlabel('Age')
ax_hex.set_ylabel('Income')

# Bubble plot
scatter = axes[1, 1].scatter(
    data['age'], data['income'], s=data['satisfaction']*50,
    c=data['satisfaction'], cmap='viridis', alpha=0.6, edgecolors='black'
)
axes[1, 1].set_title('Age vs Income (Bubble Plot)')
axes[1, 1].set_xlabel('Age')
axes[1, 1].set_ylabel('Income')
plt.colorbar(scatter, ax=axes[1, 1], label='Satisfaction')

plt.tight_layout()
plt.show()

# 3. Comparison Plots
fig, axes = plt.subplots(2, 2, figsize=(12, 8))

# Bar plot
category_counts = data['category'].value_counts()
axes[0, 0].bar(category_counts.index, category_counts.values, color='skyblue', edgecolor='black')
axes[0, 0].set_title('Category Distribution (Bar Chart)')
axes[0, 0].set_ylabel('Count')

# Grouped bar plot
grouped_data = data.groupby(['category', 'region']).size().unstack()
grouped_data.plot(kind='bar', ax=axes[0, 1], edgecolor='black')
axes[0, 1].set_title('Category by Region (Grouped Bar)')
axes[0, 1].set_ylabel('Count')
axes[0, 1].legend(title='Region')

# Stacked bar plot
grouped_data.plot(kind='bar', stacked=True, ax=axes[1, 0], edgecolor='black')
axes[1, 0].set_title('Category by Region (Stacked Bar)')
axes[1, 0].set_ylabel('Count')

# Horizontal bar plot
region_counts = data['region'].value_counts()
axes[1, 1].barh(region_counts.index, region_counts.values, color='lightcoral', edgecolor='black')
axes[1, 1].set_title('Region Distribution (Horizontal Bar)')
axes[1, 1].set_xlabel('Count')

plt.tight_layout()
plt.show()

# 4. Correlation and Heatmaps
numeric_cols = data[['age', 'income', 'education_years', 'satisfaction']].corr()

fig, axes = plt.subplots(1, 2, figsize=(14, 5))

# Correlation heatmap
sns.heatmap(numeric_cols, annot=True, fmt='.2f', cmap='coolwarm', center=0,
            square=True, ax=axes[0], cbar_kws={'label': 'Correlation'})
axes[0].set_title('Correlation Matrix Heatmap')

# Clustermap alternative
from scipy.cluster.hierarchy import dendrogram, linkage
from scipy.spatial.distance import pdist, squareform

# Create a simpler heatmap for category averages
category_avg = data.groupby('category')[['age', 'income', 'education_years', 'satisfaction']].mean()
sns.heatmap(category_avg.T, annot=True, fmt='.1f', cmap='YlGnBu', ax=axes[1],
            cbar_kws={'label': 'Average Value'})
axes[1].set_title('Average Values by Category')

plt.tight_layout()
plt.show()

# 5. Pair Plot
pair_cols = ['age', 'income', 'education_years', 'satisfaction']
plt.figure(figsize=(12, 10))
pair_plot = sns.pairplot(data[pair_cols], diag_kind='hist', corner=False)
pair_plot.fig.suptitle('Pair Plot Matrix', y=1.00)
plt.show()