12. Exploratory Data Analysis (EDA)

12.1 Question-Driven EDA

By Ravindra Bagale · Last updated: 30 September 2026

Do not plot everything. Start with questions:

1. How many rows, columns, nulls, duplicate keys?
2. What is the grain?
3. Which cities / statuses dominate?
4. Are festival days different?
5. Any impossible values (mins, amounts)?

import pandas as pd

df = pd.DataFrame({
    "order_id": [f"BLK-{i}" for i in range(1, 13)],
    "city": [
        "Pune", "Pune", "Nashik", "Nagpur", "Solapur", "Kolhapur",
        "Sambhaji Nagar", "Pune", "Nashik", "Pune", "Nagpur", "Pune",
    ],
    "amount": [64, 900, 450, 240, 1299, 58, 195, 320, 880, 110, 240, 75],
    "status": [
        "Delivered", "Delivered", "Delivered", "Delivered", "Delivered", "Cancelled",
        "Delivered", "Delivered", "Delivered", "Returned", "Delivered", "Delivered",
    ],
    "festival": [
        "None", "Diwali", "Diwali", "None", "None", "None",
        "Diwali", "Ganeshotsav", "Diwali", "None", "None", "None",
    ],
    "delivery_mins": [9, 18, 16, 11, 18, 14, 10, 12, 17, 13, 11, 8],
    "customer": [
        "Ruhi Bagale", "Amir", "Salman", "Zoya", "Raja", "Rani",
        "Shahrukh", "Ravina", "Shraddha Bagale", "Ruhi Bagale", "Amir", "Salman",
    ],
})