Traffic Predictor AI 🚦

A Python-based traffic forecasting system that uses machine learning (XGBoost) to estimate road traffic intensity based on various weather and calendar data. The project includes a data processing unit, a model training script, and a modern, dark-mode graphical user interface (GUI).

✨ Features

• Automated Data Processing: Cleaning raw traffic data and extracting key features.
• XGBoost Regression Model: High-precision traffic estimation based on historical data.
• Interactive Dashboard: User-friendly interface for adjusting prediction parameters.
• Live Weather Integration: Real-time forecast retrieval using the Open-Meteo API.
• Visual Analysis: 24-hour traffic trend visualization on a graph.

📂 Project Structure

The project consists of three main components:

1. ConverterV2.py: Data preparation script. It converts units (Kelvin -> Celsius), handles missing data, and calculates calendar variables (weekend, holiday).
2. Learning.py: Machine learning module. It loads the prepared data, trains the XGBRegressor model, and saves it into a .pkl file.
3. App.py: The main application. A customtkinter-based interface where predictions can be run using manually entered or API-fetched data.

🚀 Usage

Using the system involves three simple steps:

1. Data Processing

Run the ConverterV2.py script to transform the raw database (Metro_Interstate_Traffic_Volume.csv):

python scripts/ConverterV2.py

This will create the AI_ready_traffic_data.csv file.

2. Model Training

Train the AI model on the processed data:

python scripts/Learning.py

At the end of the process, the traffic_model(Regressor).pkl file containing the trained model will be created.

3. Launching the App

Start the graphical dashboard:

python scripts/App.py

📊 Data Source & API

• Training Data: UCI Machine Learning Repository - Metro Interstate Traffic Volume
• Weather Forecast: Open-Meteo API (No API key required)

⚖️ License

It is free to use and modify.