aabusoba/raqib-model-ai

Raqib AI is a dual-model Traffic Intelligence Hub designed for Libya. It utilizes advanced machine learning (XGBoost) and data analysis to predict road accident severity and provide actionable safety recommendations. # 🚦 Raqib AI - Traffic Decision Support System (DSS) **Live Demo:** raqib-ai.nafusah.tech Raqib AI is a dual-model Traffic Intelligence Hub designed for Libya. It utilizes advanced machine learning (XGBoost) and data analysis to predict road accident severity and provide actionable safety recommendations. ## 🚀 Quick Start - Docker (Recommended) The easiest way to run the entire system (both models + API + Gateway) is using Docker Compose. ### Prerequisites - Docker and Docker Compose installed. ### Installation 1. Clone the repository. 2. Build and start the containers: ```bash docker-compose up -d --build ``` 3. Access the system at: - **Main System Status (JSON):** localhost - **Model Design 1 Dashboard:** localhost - **Model Design 2 Dashboard:** localhost --- ## 🛠 Manual Local Installation If you prefer to run the models independently without Docker: ### Prerequisites - Python 3.9+ installed. ### Steps for each model (Design 1 or Design 2) 1. **Navigate to the model folder:** ```bash cd model_desgin_1 # or model_desgin_2 ``` 2. **Install dependencies:** ```bash pip install -r requirements.txt ``` 3. **Train the AI Models (Mandatory for the first time):** This will sync data from Kaggle and generate weights (`.pkl`) and stats summary (`.json`). ```bash python train.py ``` 4. **Launch the Dashboard (Streamlit):** ```bash streamlit run app.py ``` 5. **Launch the API (Flask):** ```bash python api.py ``` --- ## 📊 Feature Highlights - **Model Design 1:** Focuses on statistical classification and classic ML. - **Model Design 2:** Advanced XGBoost model integrating vehicle and driver data. - **Lightweight Mode:** Once trained, the system runs using model weights (`.pkl`) and a tiny JSON summary, requiring NO local CSV storage. - **Auto-Sync:** Real-time data synchronization from Kaggle HUB. ## 🔗 Routing Architecture The system uses an Nginx gateway to route traffic: - `/mod1/` -> Dash …