🛒 E-Commerce Customer Analytics & Revenue Optimization Engine

An end-to-end data processing, exploratory data analysis (EDA), and customer profiling pipeline applied to a real-world transnational dataset containing 541,909 raw transaction records.

---

📌 Executive Summary

Modern e-commerce enterprises require granular insights into customer purchasing behaviors to optimize target marketing and minimize churn. This project builds a robust data engineering and analytics pipeline that:

1. Cleans and transforms noisy operational transaction data (handling cancellations, unit pricing anomalies, and missing identities).
2. Engineers temporal and financial metrics (RFM readiness, transaction totals, time features).
3. Extracts high-value business insights across 3,665 product SKUs, 4,338 unique customers, and 37 countries.

---

📊 Key Pipeline Findings

• Data Quality & Retention: Processed and cleaned 541,909 raw entries down to 392,692 high-fidelity records (72.46% retention rate) after removing 149,217 noisy rows (cancellations, negative/zero prices, missing CustomerID, exact duplicates).
• Total Revenue Analyzed: £8,887,208.89 generated across 18,532 unique invoices.
• Core Metrics Summary:
  • Basket Quantity: Mean = 9.55 units (Median = 3.0 units).
  • Unit Price: Mean = £4.61 (Median = £2.08).
  • Order Value: Mean = £22.63 (Median = £9.90).

---

🏗 Repository Architecture

ecommerce-customer-analytics-pipeline/
├── notebooks/
│   └── customer_segmentation_eda.ipynb   # Complete ETL, Cleaning & Visualization Pipeline
├── docs/
│   └── project_report.pdf                 # Executive Summary & Technical Presentation
├── .gitignore                             # Python gitignore configuration
├── LICENSE                                # MIT License
└── README.md                              # Technical Project Overview