Exploratory Data Analysis

What I do

• Investigate datasets systematically
• Identify patterns, trends, and anomalies
• Generate hypotheses from data
• Create summary visualizations
• Understand variable relationships
• Assess data quality and requirements

When to use me

Use me when:

• Starting any data science project
• Understanding new datasets
• Preparing for modeling
• Answering ad-hoc analytical questions
• Discovering insights in data

Key Concepts

EDA Workflow

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

# 1. Data Overview
df = pd.read_csv("data.csv")
print(f"Shape: {df.shape}")
print(f"Columns: {df.columns.tolist()}")
print(df.info())

# 2. Summary Statistics
print(df.describe(include="all"))

# 3. Missing Values
missing = df.isnull().sum()
missing_pct = (missing / len(df)) * 100
print(pd.DataFrame({"Missing": missing, "%": missing_pct}))

# 4. Univariate Analysis
for col in df.select_dtypes(include=[np.number]).columns:
    fig, ax = plt.subplots(1, 2, figsize=(10, 4))
    df[col].hist(ax=ax[0], bins=30)
    df.boxplot(column=col, ax=ax[1])

# 5. Bivariate Analysis
numeric_cols = df.select_dtypes(include=[np.number]).columns
corr = df[numeric_cols].corr()

# 6. Categorical Analysis
for col in df.select_dtypes(include="object").columns:
    print(df[col].value_counts().head(10))

Visual EDA Techniques

• Histograms: Distribution shape
• Box plots: Outliers and quartiles
• Scatter plots: Relationships
• Heatmaps: Correlation
• Pair plots: Multiple relationships
• Count plots: Categorical counts