In today’s hyperconnected world, choosing the right data plan can be overwhelming. Most people settle for what’s available without knowing if it’s cost-efficient or tailored to their needs.
That’s where DataPlan Recommender steps in — a machine learning-powered platform that personalizes data plan suggestions based on individual usage behavior and demographics.
🧠 What is DataPlan Recommender?
DataPlan Recommender is a smart system that leverages CatBoost and XGBoost machine learning algorithms to recommend the most suitable and cost-effective data plans. It analyzes user data including demographics, occupation, location, and usage patterns to ensure that users get personalized and optimized suggestions.
The core goal? Reduce unnecessary expenses and ensure uninterrupted, value-driven internet access.
🗂 Dataset Overview
The project uses a rich dataset with diverse attributes that mirror real-world telecom usage scenarios:
👤 Personal Info: Name, Age, Mobile Number, Password
💼 Occupation: IT Professionals, Doctors, Teachers, etc.
📍 Location: Users from various cities
📊 Usage: Daily & Monthly Internet Usage (GB/MB)
📱 Devices: Smartphone, Tablet, Smart TVs
🌐 Network Preferences: 4G/5G, SIM Type (Airtel, VI, BSNL)
📦 Package Details: Data Limit, Package Cost, and Network Plan
This data helps identify usage trends and predict the most fitting data plan for each user.
Methodology:
The methodology involves a structured machine learning pipeline:
- Data Preprocessing – Clean, normalize, and format user data
- Feature Engineering – Extract insights like usage trends, cost per GB
- Model Training – Use CatBoost/XGBoost to train prediction models
- Backend Integration – Flask APIs to serve predictions
- Frontend Interaction – Responsive UI with real-time suggestions
📊 The system learns from historical data to uncover the best match between user profile and available data plans.
💻 Tech Stack Here’s what powers the application:
📈 Results The application delivers effective and insightful outcomes:
✅ Prediction Accuracy: High accuracy (0.98) in predicting user needs using advanced models
🙌 User Engagement: Intuitive front-end helps users compare and choose plans easily
💸 Optimized Costs: Suggests cost-efficient plans based on actual usage
📡 Telco Insights: Provides feedback to network providers for better pricing models
🖼 Snapshots Here are some glimpses of the platform in action:
🎥 Demo See the recommender system in action!
Github repo link:
🚀 What’s Next?
I'm working on:
- Adding more user behavioral data like app usage patterns
- Supporting real-time plan updates via telecom APIs
- Creating a mobile-first version of the platform
- Integrating with billing portals for one-click plan purchases
💬 Final Thoughts The DataPlan Recommender showcases how AI/ML can transform something as mundane as choosing a mobile data plan into a smart, personalized, and cost-efficient process. Whether you’re a student, a working professional, or a telco provider, this tool offers something valuable.
Top comments (1)
Good Effort