🛒 E-Commerce Supply Chain Analytics Project

End-to-End Modern Data + Analytics + ML Project

A production-style analytics platform built using dbt + Microsoft Fabric + SQL + Python + Machine Learning + Power BI, designed to answer real business questions and drive data-backed decisions.

---

📌 Project Overview

This project simulates a real-world e-commerce supply chain analytics system, starting from raw transactional data and ending with business-ready dashboards and ML-driven customer intelligence.

The system answers:

• Who are our customers?
• Who generates revenue?
• Who is risky?
• Who should we retain, fix, or ignore?
• What operational issues impact revenue?

---

🧠 Business Philosophy

Layer	Answers
Raw / Bronze	What data do we have?
Silver	What happened?
Gold + ML	What should we do?
Power BI	How should the business act?

This dashboard is decision-oriented, not number-oriented.

---

🏗️ Architecture (End-to-End)

---

📂 Project Folder Structure

E-COMMERCE SUPPLY CHAIN ANALYTICS PROJECT
├── assets/
│   ├── E commerce Sales Analytics.pbix
│   ├── E Commerce Supply Chain Analytics.pdf
│   └── Project Pipeline.png
├── ecommerce_fabric/               # Main dbt Project Folder
│   ├── analyses/
│   ├── logs/
│   ├── macros/
│   ├── models/
│   │   ├── bronze/                 # Raw data standardization
│   │   ├── gold/                   # Aggregated business entities
│   │   ├── silver/                 # Feature engineering & cleaning
│   │   ├── staging/                # Initial source casting
│   │   ├── schema.yml
│   │   └── sources.yml
│   ├── seeds/
│   ├── snapshots/
│   ├── tests/
│   │   └── silver/
│   ├── .gitkeep
│   ├── .gitignore
│   ├── dbt_project.yml
│   ├── package-lock.yml
│   ├── packages.yml
│   └── README.md
├── notebooks/                      # Data Science & ML workflow
│   ├── gold_customer_ml_features.csv
│   └── ml_customer_insights.ipynb
├── venv/                           # Python Virtual Environment
├── index.html                      # Project Landing Page
├── README.md                       # Main Documentation
└── requirements.txt                # Python Dependencies

---

🥉 Bronze Layer (Raw Standardization)

Purpose: Clean & standardize raw data without business assumptions.

Models

• stg_customers
• stg_orders
• stg_payments
• stg_returns
• stg_shipments

---

🥈 Silver Layer (Feature Engineering)

Purpose: Explain what happened using analytical features.

Core Models

• dim_customers
• fact_orders_clean
• customer_order_features
• customer_payment_features
• customer_return_features
• customer_fulfillment_features
• silver_customer_features

---

🤖 Machine Learning Layer

Notebook: customer_segmentation_ml.ipynb

Techniques

• Feature scaling
• One-hot encoding
• KMeans clustering
• PCA visualization
• Cluster profiling

---

🥇 Gold Layer (Business Ready)

Models

• gold_customer_360
• gold_customer_segments
• gold_customer_ml_features

📊 Power BI Dashboard

Pages

1. Executive Overview
2. Customer Segment Intelligence
3. Risk & Loss Analytics
4. Customer 360
5. Business Actions

Download the Full Dashboard (PDF)

---

🧪 Data Quality & Testing

• 60+ dbt tests
• Not-null, uniqueness, accepted values
• Rate sanity checks
• ML output validation

🚀 Key Outcomes

• Identified risky and valuable customers
• Quantified revenue leakage
• Enabled ML-driven segmentation
• Built executive-level dashboards

---

👨‍💻 Author

Prince Singh
Data Analytics & Data Science
India

---