AI-powered customer intelligence platform built using Machine Learning, Behavioral Analytics, and Power BI.
Understanding customer behavior is critical for improving customer retention, increasing revenue, and delivering personalized experiences.
This project applies Machine Learning-based Customer Segmentation and Behavioral Analytics on retail transaction data to identify customer personas, analyze purchasing patterns, and generate actionable business insights.
The solution combines:
✅ RFM Analysis
✅ Feature Engineering
✅ PCA (Dimensionality Reduction)
✅ K-Means Clustering
✅ DBSCAN Clustering
✅ Behavioral Analytics
✅ Interactive Power BI Dashboards
Businesses often struggle to answer questions such as:
- Who are our most valuable customers?
- Which customers are likely to become inactive?
- How do different customer groups behave?
- Which segments respond to promotions?
- How can marketing campaigns be personalized?
This project addresses these challenges using data-driven customer intelligence.
The dataset contains:
- Customer Information
- Product Categories
- Purchase Quantity
- Pricing Information
- Discounts
- Payment Methods
- Transaction History
| Metric | Value |
|---|---|
| Records | 100,000+ |
| Customers | 95,000+ |
| Features | 10+ |
- Dataset Exploration
- Data Quality Validation
- Missing Value Checks
- Data Type Verification
| Feature | Description |
|---|---|
| Recency | Days since last purchase |
| Frequency | Number of purchases |
| Monetary | Customer spending value |
| Feature | Description |
|---|---|
| AvgQuantity | Average quantity purchased |
| AvgDiscount | Average discount received |
| PreferredPayment | Most used payment method |
| FavoriteCategory | Most purchased category |
| AvgPurchaseInterval | Average time between purchases |
Used PCA to:
- Reduce dimensionality
- Remove feature redundancy
- Improve clustering performance
- Enhance visualization
Applied K-Means clustering to identify customer personas.
| Segment | Description |
|---|---|
| ⭐ Premium Loyal Customers | High-value customers with strong spending behavior |
| 👥 Regular Customers | Consistent and active customer group |
| 🔄 Inactive Customers | Low-engagement customers with retention potential |
Used DBSCAN to:
- Identify behavioral outliers
- Detect unusual customer patterns
- Validate customer segmentation
📌 Active Customers
📌 Customer Value
📌 Repeat Purchase Rate
📌 Discount Sensitivity
📌 High-Value Customers
- Revenue Contribution by Segment
- Customer Activity & Retention Risk
- Segment Distribution Analysis
- Customer Value Analysis
📌 Average Basket Size
📌 Digital Payment Adoption
📌 Promotional Purchase Rate
📌 Average Purchase Gap
- Product Preference Analysis
- Payment Behavior Analysis
- Purchase Quantity Analysis
- Customer Persona Comparison
- Segment Performance Analysis
✅ Three distinct customer personas were identified using Machine Learning.
✅ Regular Customers represent the largest customer segment.
💰 Regular Customers contribute the highest customer value.
💰 Premium Loyal Customers maintain strong spending despite lower purchase frequency.
💰 Inactive Customers generate the lowest revenue contribution.
🛒 Customers purchase an average of 5 products per transaction.
💳 Digital payment adoption exceeds 70%.
🏷️ Promotions influence approximately 10% of purchasing behavior.
📦 Product category preferences vary significantly across customer segments.
📦 Different customer groups exhibit distinct purchasing patterns.
🎯 Inactive Customers represent the biggest retention opportunity.
🎯 Repeat purchasing behavior is strongest among Regular Customers.
- Re-engage inactive customers through targeted campaigns.
- Monitor customers showing declining engagement.
- Reward Premium Loyal Customers.
- Introduce personalized loyalty benefits.
- Create segment-specific promotions.
- Personalize product recommendations.
- Expand digital payment incentives.
- Encourage cash users to adopt digital channels.
- Python
- Pandas
- NumPy
- Scikit-Learn
- PCA
- K-Means
- DBSCAN
- Matplotlib
- Seaborn
- Power BI
advanced-customer-segmentation-ml/
├── dashboard/
│ └── Customer_Segmentation.pbix
├── images/
│ ├── 01_customer_intelligence_dashboard.png
│ ├── 02_behavioral_analytics_dashboard.png
│ └── 03_key_findings_dashboard.png
│ └── project_workflow.png
├── notebooks/
│ ├── Data Cleaning
│ ├── Feature Engineering
│ ├── PCA Analysis
│ ├── K-Means Clustering
│ └── DBSCAN Analysis
├── reports/
│ ├── Project Summary
│ └── Final Report
├── README.md
├── requirements.txt
├── LICENSE
└── .gitignore
| Notebook | Purpose |
|---|---|
| 01_data_understanding | Dataset exploration |
| 02_rfm_feature_engineering | Feature engineering |
| 03_pca_preprocessing | PCA implementation |
| 04_kmeans_clustering | Customer segmentation |
| 05_customer_persona_analysis | Persona creation |
| 06_dbscan_clustering | Outlier analysis |
| 07_dashboard_visualization | Dashboard preparation |
This project demonstrates:
✅ Data Cleaning & Preparation
✅ Feature Engineering
✅ Machine Learning
✅ Customer Segmentation
✅ Behavioral Analytics
✅ Business Intelligence
✅ Power BI Dashboarding
✅ Business Storytelling
Bala Venkatesh G
Big Data Engineer | Analytics Enthusiast
⭐ If you found this project useful, consider giving it a star!