Raiffeisen Bank Customer Analytics Dashboard

Overview

A sophisticated customer analytics dashboard built for Raiffeisen Bank that leverages machine learning to segment customers and provide personalized product recommendations. The application uses K-means clustering to categorize customers into three distinct segments, enabling targeted marketing strategies and improved customer retention.

Hackathon Achievement

This project was developed during the 24-hour Data Hackathon organized by Raiffeisen Bank in 2024. Our team won first place, competing against 5 other teams. The project is intended for hackathon demonstration purposes only.

Contributors

• Artin Rexhepi
• Daris Dragusha

Live Demo

Try out the live demo of our application here: Raiffeisen Bank Analytics Dashboard

Features

1. Customer Segmentation

• Real-time customer segmentation using machine learning
• Three distinct customer segments:
  • Engaged Mid-Tier Customers
  • Low-Balance Loyalists
  • High-Balance At-Risk Customers
• Detailed segment analysis and characteristics

2. Interactive Dashboard

• Overview section with key metrics and visualizations
• Customer profiling with instant segment prediction
• Product recommendations tailored to each segment
• Comprehensive data visualization using Plotly

3. Email Marketing Integration

• Automated product recommendation emails
• Personalized HTML email templates
• Direct customer communication capabilities

Technology Stack

• Python 3.11+
• Streamlit for web interface
• Scikit-learn for machine learning
• Pandas for data manipulation
• Plotly for interactive visualizations
• SMTP for email functionality

Project Structure

• ui.py: Main application interface and logic
• modelPrediction.ipynb: Jupyter notebook containing model development
• clustered_data.csv: Processed customer data with cluster assignments
• scaler_and_model.pkl: Saved ML model and scaler
• requirements.txt: Project dependencies

Customer Segments

Engaged Mid-Tier Customers

• Middle-aged customers
• Low tenure, average balance
• Moderate activity with 1-2 products
• Slight exit risk

Low-Balance Loyalists

• Middle-aged customers
• Moderate tenure, low balances
• Multiple products
• High salary, strong loyalty

High-Balance At-Risk Customers

• Middle-aged customers
• Moderate tenure
• Highest average balance
• Fewer products, highest exit risk