Internet package recommender for WE users in Egypt. Analyzes user internet usage behavior to recommend the most suitable data plan , helping avoid unexpected overages and extra costs.
# PacketMatch — WE Package Recommendation System
---
## Project Overview
PacketMatch is an intelligent Machine Learning-powered recommendation system designed to help WE internet users select the most suitable internet package based on their real usage behavior and consumption patterns.
The project addresses a real-world problem faced by many internet users in Egypt: internet bundles running out unexpectedly before month-end, resulting in extra costs and poor user experience.
The system analyzes user behavior such as:
- Daily internet consumption
- Number of connected devices
- Usage type
- Remaining quota
- Billing cycle information
Then predicts the optimal WE internet package using a trained XGBoost classification model.
The project combines Web Scraping, Synthetic Data Generation, Feature Engineering, Machine Learning, and Interactive Streamlit Deployment.
---
## Problem Statement
Many WE internet users in Egypt struggle with:
- Internet packages running out before the end of the month
- No clear visibility into where data consumption is going
- Unexpected spikes in usage leading to confusion and frustration
- Paying unnecessary extra bundle costs without understanding why
- Difficulty choosing the right package from available options
**Users are not struggling with lack of data plans —
they are struggling with lack of visibility and control over their internet usage.**
---
## Proposed Solution
PacketMatch transforms raw internet usage data into personalized package recommendations using Machine Learning.
The system:
- Analyzes user internet behavior
- Detects inefficient consumption patterns
- Predicts the most suitable package
- Helps reduce extra bundle costs
- Supports smarter internet usage decisions
---
## Project Structure
```
PacketMatch/
│
├── data/
│ ├── WE_centrals.xlsx
│ ├── WE_Dataset.csv
│ └── we_plans.xlsx
|
├── data_generation/
│ └── WE_Dataset_Generator_v6.py
|
├── images/
│ ├── Package Information. …