hadsa129/customer-behavior-analysis-orange-tunisia

# Customer Behavior Analysis & Personalized Recommendation Platform - Orange Tunisia ## 🎯 Objective This advanced analytics platform enables understanding of Orange Tunisia customer behavior and provides personalized recommendations. It aims to design an intelligent platform for analyzing Orange Tunisia customer behaviors, predicting churn, and automatically recommending personalized offers, particularly through the Max It application. It aligns with the operator's digital transformation strategy, focusing on marketing personalization and customer engagement optimization.It combines big data, machine learning, and AI techniques to deliver actionable insights to marketing teams. --- ## 📌 Project Objectives ### Business Objectives - Analyze multi-channel behaviors (Max It, USSD...) - Dynamically segment customers - Predict churn - Recommend personalized offers - Generate adapted marketing messages - Assist marketing teams with interactive dashboards ### Technical Objectives - Distributed Big Data architecture - AI models for segmentation, churn, recommendation - LLM integration for marketing content generation - Dynamic decision-making dashboards ## 🏗️ System Architecture The following diagram illustrates the complete architecture of the Customer Behavior Analysis & Personalized Recommendation Platform, including the Big Data ETL pipeline, AI/ML components, Django backend services, deployment infrastructure, and user interfaces. ### Architecture Overview - **Data Sources**: Customer profiles, transactions, options history, and external APIs - **Big Data & ETL**: Apache Spark, HDFS, Hive Metastore, distributed processing - **AI & ML Core**: - Customer Segmentation (K-Means) - Churn Prediction (XGBoost) - Recommendation Engine - LLM-based marketing generation using Qwen + LangChain - **Backend Services**: - Django REST Framework - PostgreSQL - Redis cache - **Deployment & Infrastructure**: - Docker - Nginx + Gunicorn - **User Interfaces**: - Marketing Dashbo …