Exploratory Data Analysis (EDA)

Overview

Exploratory Data Analysis (EDA) is the critical first step in data science projects, systematically examining datasets to understand their characteristics, identify patterns, and assess data quality before formal modeling.

Core Concepts

• Data Profiling: Understanding basic statistics and data types
• Distribution Analysis: Examining how variables are distributed
• Relationship Discovery: Identifying patterns between variables
• Anomaly Detection: Finding outliers and unusual patterns
• Data Quality Assessment: Evaluating completeness and consistency

When to Use

• Starting a new dataset analysis
• Understanding data before modeling
• Identifying data quality issues
• Generating hypotheses for testing
• Communicating insights to stakeholders

Implementation with Python

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

# Load and explore data
df = pd.read_csv('customer_data.csv')

# Basic profiling
print(f"Shape: {df.shape}")
print(f"Data types:\n{df.dtypes}")
print(f"Missing values:\n{df.isnull().sum()}")
print(f"Duplicates: {df.duplicated().sum()}")

# Statistical summary
print(df.describe())
print(df.describe(include='object'))

# Distribution analysis - numerical columns
fig, axes = plt.subplots(2, 2, figsize=(12, 8))
df['age'].hist(bins=30, ax=axes[0, 0])
axes[0, 0].set_title('Age Distribution')

df['income'].hist(bins=30, ax=axes[0, 1])
axes[0, 1].set_title('Income Distribution')

# Box plots for outlier detection
df.boxplot(column='age', by='region', ax=axes[1, 0])
axes[1, 0].set_title('Age by Region')

# Categorical analysis
df['category'].value_counts().plot(kind='bar', ax=axes[1, 1])
axes[1, 1].set_title('Category Distribution')
plt.tight_layout()
plt.show()

# Correlation analysis
numeric_df = df.select_dtypes(include=[np.number])
correlation_matrix = numeric_df.corr()

plt.figure(figsize=(10, 8))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', center=0)
plt.title('Correlation Matrix')
plt.show()

# Multivariate relationships
sns.pairplot(df[['age', 'income', 'education_years']], diag_kind='hist')
plt.show()

# Skewness and kurtosis
print("\nSkewness:")
print(numeric_df.skew())
print("\nKurtosis:")
print(numeric_df.kurtosis())

# Percentile analysis
print("\nPercentiles for Age:")
print(df['age'].quantile([0.25, 0.5, 0.75, 0.95, 0.99]))

# Missing data patterns
missing_pct = (df.isnull().sum() / len(df) * 100)
missing_pct[missing_pct > 0].sort_values(ascending=False)

# Value count analysis
print("\nCustomer Types Distribution:")
print(df['customer_type'].value_counts(normalize=True))

# Advanced EDA: Groupby analysis
print("\nGroupBy Analysis:")
print(df.groupby('region')[['age', 'income']].agg(['mean', 'median', 'std']))

# Correlation with target variable
if 'target' in df.columns:
    target_corr = df.corr()['target'].sort_values(ascending=False)
    print("\nFeature Correlation with Target:")
    print(target_corr)

# Data type breakdown
print("\nData Type Summary:")
print(df.dtypes.value_counts())

# Unique value count
print("\nUnique Value Counts:")
print(df.nunique().sort_values(ascending=False))

# Variance analysis
print("\nVariance per Feature:")
numeric_cols = df.select_dtypes(include=[np.number]).columns
for col in numeric_cols:
    variance = df[col].var()
    print(f"  {col}: