🛒 E-Commerce Data Analytics & Predictive Insights Project

🔖 Overview

This project demonstrates a full-cycle e-commerce data analysis solution developed collaboratively using modern data technologies.

As part of the Power BI team, my key role was customer segmentation—analyzing customer behavior to derive actionable insights that inform marketing strategies and product recommendations. The project was organized and tracked through Notion, with clear task delegation and transparent collaboration.

---

🔗 Project Artifacts

• 📎 Notion (team documentation)
• 📊 Power BI Dashboards
• 📁 SQL Scripts and Python Notebooks

---

🧩 Tools & Technologies

• ☁️ Cloud Platform: Azure SQL Database
• 🛠 Data Cleaning & Processing: SQL Server, Python (pandas, pyodbc, Mito)
• 📊 Visualization: Power BI
• 🧠 Machine Learning: scikit-learn, XGBoost
• 💬 NLP Querying: Vanna AI

---

🗂️ Project Phases

☁️ 1. Creating & Connecting Azure SQL Database

• Set up and configured a cloud-hosted SQL database for collaborative access.
• Managed the secure connection between local tools (Power BI, Python) and the Azure SQL instance.

---

🔍 2. Data Cleaning & Validation (SQL Server)

• Performed rigorous audits using dynamic SQL:

  • Removed duplicates and invalid rows
  • Standardized date and numeric formats
  • Detected anomalies:
    • Orders with zero/negative values
    • Inventory with zero quantity
    • Customers with no orders
    • Overstocked or unused products

• 💡 Created secure, department-specific views for role-based access

---

📈 3. Business Intelligence Dashboard (Power BI)

🎯 My Role: Customer Segmentation Analysis
I was responsible for identifying customer groups using behavioral data (e.g., frequency of purchase, total spend, return behavior) and visualizing patterns via interactive dashboards.

📊 Key KPIs:

• 🧾 Total Revenue: $45M (+24.6%)
• 📦 Total Orders: 10.2K (+24.57%)
• 🔁 Return Rate: 1% (excellent satisfaction)

---

🔍 Insights Delivered:

• 📌 Top-performing customer segments
• 📆 Monthly ordering trends and anomalies
• 🔁 Return & shipping behavior analysis
• ⚠️ Inventory level alerts

---

💡 Recommendations Based on Dashboards:

• 🏅 Implement loyalty programs for high-retention customer segments
• 🔄 Adjust return policies for high-return products
• 🎯 Optimize marketing strategies for under-engaged groups
• 🧺 Bundle frequently bought-together products
• 🚨 Act on inventory alerts for out-of-stock and overstocked items

---

💬 4. Natural Language SQL (Vanna AI)

• Enabled natural language querying for non-technical users
• Translated business questions into real-time SQL queries
• Enhanced team collaboration and faster insight delivery

---

🧠 5. Machine Learning Pipeline (Python)

• Built ML models to predict customer order value using:
  • Random Forest Regressor (baseline)
  • XGBoost Regressor (final model)
• Used Mito, pandas, pyodbc, scikit-learn, and joblib
• Achieved high accuracy in predicting repeat customer behavior

📈 Outcome: Informed personalized offers and targeted marketing strategies based on predicted customer lifetime value.

---

🎯 Final Insights & Recommendations

• ✅ 91% 30-day customer retention rate → Launch tiered loyalty programs
• ✅ Products rated ≤ 3 stars → Launch feedback forms and review incentives
• ✅ Frequently bought-together pairs → Bundle for upselling
• ✅ Underused discounts → Revisit promotion visibility & targeting

---

🏆 My Key Contributions

• 🎯 Specialized in Customer Segmentation using Power BI
• 📊 Designed and implemented dashboards with actionable KPIs
• 🧠 Transformed raw customer data into impactful insights
• 🤝 Collaborated with SQL & ML engineers in a cross-functional team
• 💡 Contributed to final project presentation and competition entry

---

📬 Let’s Connect

For feedback, collaboration, or inquiries about the full report or dashboards:

• 📧 Email: hasnaaahmed745@gmail.com
• 💼 LinkedIn: [linkedin.com/in/hasnaaahmed]([https://www.linkedin.com/in/hasnaaahmed]

---