Customer Behavior Analysis

A full end-to-end data analysis project analyzing customer shopping behavior using Python, SQL, and Power BI.

Project Overview

This project explores a customer shopping behavior dataset to uncover insights about revenue, customer segments, discount usage, product preferences, and subscription patterns.

Tools & Technologies

• Python (pandas) — data cleaning and feature engineering
• MySQL — data storage and analytical queries
• Power BI — interactive dashboard and data visualization

Dataset

Customer shopping behavior dataset containing 3,900 records with fields including:

• Customer demographics (age, gender)
• Purchase details (amount, category, item)
• Behavioral data (subscription status, discount usage, frequency of purchases, review ratings)

What I Did

Python (Data Cleaning & Feature Engineering)

• Handled missing values using category-wise median imputation for review ratings
• Standardized column names (lowercase, underscores)
• Created new features:
  • age_group — segmented customers into Young Adult, Adult, Middle Age, Senior using quartile binning
  • purchase_frequency_days — converted purchase frequency labels into numeric days
• Removed redundant columns after validation
• Exported cleaned data to MySQL database using SQLAlchemy

SQL (Data Analysis — 10 Questions)

• Total revenue by gender
• High-spending discount users vs average
• Top 5 products by average review rating
• Shipping type vs average purchase amount
• Subscriber vs non-subscriber spend comparison
• Top 5 products by discount rate
• Customer segmentation — New, Returning, Loyal (using CTE)
• Top 3 products per category (using Window Functions)
• Repeat buyer subscription analysis
• Revenue contribution by age group

Power BI (Dashboard)

• KPI cards: Total Customers, Average Purchase Amount, Average Review Rating
• % of Customers by Subscription Status (donut chart)
• Revenue by Category (bar chart)
• Revenue by Gender (bar chart)
• Revenue by Age Group (horizontal bar chart)
• Average Rating by Product — Top 5 (bar chart)
• Interactive slicers: Subscription Status, Gender, Category, Shipping Type

Key Insights

• Clothing is the highest revenue-generating category
• Young Adults contribute the most revenue by age group
• 73% of customers are not subscribed — opportunity for growth
• Subscribed customers spend more on average than non-subscribers
• Discount usage does not significantly reduce average spend

Project Structure

customer-behavior-analysis/
│
├── customer_behavior.ipynb        # Python data cleaning & feature engineering
├── customer_behavior.sql.ipynb    # SQL analytical queries
└── customer behavior dashboard.pbix  # Power BI dashboard

How to Run

1. Clone this repository
2. Open customer_behavior.ipynb in Jupyter Notebook
3. Run all cells to clean the data and push to MySQL
4. Open customer_behavior.sql.ipynb to run SQL queries
5. Open the .pbix file in Power BI Desktop to view the dashboard

Author

Kartik — Aspiring Data Analyst

• GitHub: Karitik