Vehicle Severity Prediction - Django & Machine Learning

📋 Project Overview

This Django-based Machine Learning application predicts vehicle accident severity using a trained classification model. The system analyzes various accident-related factors such as speed, weather conditions, road type, and traffic density to categorize accidents into Minor, Major, or Fatal severity levels. The ML model is seamlessly integrated into a Django web application with both web interface and REST API support.

🚗 Demo Video

ml_Django_project.1.mp4

Complete demonstration of vehicle accident severity prediction system!

✨ Key Features

Core Functionality

• 🎯 Accident Severity Prediction - ML-powered classification system
• 📊 Multi-Factor Analysis - Considers 6 key accident parameters
• 🌐 Web Interface - User-friendly form for input data
• 🔗 REST API - JSON-based prediction endpoints
• 📱 Responsive Design - Mobile-friendly interface
• ⚡ Real-Time Predictions - Instant severity assessment

Machine Learning Features

• Classification Model - Trained on historical accident data
• Feature Engineering - Optimized input parameter encoding
• Model Serialization - Joblib-based model persistence
• Prediction Categories - Three-tier severity classification
• Input Validation - Robust data preprocessing

🛠️ Technology Stack

Technology	Version	Purpose
Django	5.x	Web Framework
Python	3.x	Backend Language
Scikit-Learn	1.x	Machine Learning Library
Joblib	Latest	Model Serialization
NumPy	Latest	Numerical Computing
Pandas	Latest	Data Processing
Bootstrap	5.x	UI Framework
HTML/CSS/JS	Latest	Frontend Technologies

🧠 Machine Learning Model

Input Parameters

The prediction model analyzes the following accident factors:

1. Speed (Continuous Variable)

• Type: Integer (km/hr)
• Range: 0-200 km/hr
• Impact: Higher speeds typically increase severity

2. Weather Condition (Categorical)

Code	Condition	Description
1	Clear	Optimal visibility and road conditions
2	Rainy	Reduced visibility and slippery roads
3	Fog	Severely limited visibility

3. Road Type (Categorical)

Code	Type	Characteristics
1	Highway	High-speed, controlled access roads
2	City Road	Urban streets with traffic signals
3	Rural Road	Less maintained, narrow roads

4. Time of Day (Categorical)

Code	Period	Risk Factors
1	Morning	Rush hour traffic, commuter patterns
2	Afternoon	Peak traffic, good visibility
3	Night	Reduced visibility, fatigue factors

5. Vehicle Type (Categorical)

Code	Type	Vulnerability Level
1	Car	Moderate protection
2	Motorcycle	High vulnerability
3	Truck	Heavy vehicle impact

6. Traffic Density (Categorical)

Code	Density	Impact on Accidents
1	Low	Higher speeds, fewer obstacles
2	Medium	Moderate congestion
3	High	Stop-and-go traffic, multi-vehicle risks

Prediction Categories

The model classifies accidents into three severity levels:

Severity	Description	Characteristics
Minor	Low-impact accidents	Property damage, minor injuries
Major	Moderate-impact accidents	Significant injuries, hospitalization
Fatal	High-impact accidents	Life-threatening or fatal outcomes

🚀 Installation & Setup

Prerequisites

Ensure you have the following installed:

• Python 3.8+
• pip (Python package manager)
• Git
• Virtual Environment (recommended)

Step-by-Step Installation

1. Clone the Repository

git clone https://github.com/ARONAGENT/vehicle-severity-ml.git
cd vehicle-severity-ml

2. Create Virtual Environment

# Create virtual environment
python -m venv mlproject

# Activate virtual environment
# Windows:
mlproject\Scripts\activate

# macOS/Linux:
source mlproject/bin/activate

3. Install Dependencies

# Core dependencies
pip install django
pip install scikit-learn
pip install joblib
pip install numpy
pip install pandas

# Additional packages
pip install matplotlib  # For data visualization
pip install seaborn     # For advanced plotting
pip install requests    # For API testing

4. Create Django Project Structure

# Create Django project
django-admin startproject vehicleSeverity
cd vehicleSeverity

# Create Django app
python manage.py startapp ml_predictor

5. Configure Django Settings

Update settings.py:

INSTALLED_APPS = [
    'django.contrib.admin',
    'django.contrib.auth',
    'django.contrib.contenttypes',
    'django.contrib.sessions',
    'django.contrib.messages',
    'django.contrib.staticfiles',
    'ml_predictor',  # Add your ML app
]

# Static files configuration
STATIC_URL = '/static/'
STATICFILES_DIRS = [BASE_DIR / "static"]

# Media files for model storage
MEDIA_URL = '/media/'
MEDIA_ROOT = BASE_DIR / "media"

6. Run Database Migrations

python manage.py makemigrations
python manage.py migrate

7. Start Development Server

python manage.py runserver

8. Access the Application

Visit http://127.0.0.1:8000/ to use the prediction interface.

📁 Project Structure

🔗 API Documentation

REST API Endpoints

1. Predict Accident Severity

POST /api/predict/
Content-Type: application/json

Request Body:

{
    "speed": 80,
    "weather": 2,
    "road_type": 1,
    "time_of_day": 3,
    "vehicle_type": 1,
    "traffic_density": 3
}

Response:

{
    "severity": "Major",
    "confidence": 0.85,
    "factors": {
        "speed": "High risk factor",
        "weather": "Moderate risk",
        "night_time": "Increased risk"
    },
    "recommendation": "Exercise extreme caution"
}

2. Get Model Information

GET /api/model-info/

Response:

{
    "model_type": "Random Forest Classifier",
    "accuracy": 0.89,
    "features": [
        "speed", "weather", "road_type", 
        "time_of_day", "vehicle_type", "traffic_density"
    ],
    "categories": ["Minor", "Major", "Fatal"],
    "last_trained": "2024-01-15"
}

3. Batch Predictions

POST /api/predict-batch/
Content-Type: application/json

Request Body:

{
    "predictions": [
        {
            "speed": 60,
            "weather": 1,
            "road_type": 2,
            "time_of_day": 1,
            "vehicle_type": 1,
            "traffic_density": 2
        },
        {
            "speed": 120,
            "weather": 3,
            "road_type": 1,
            "time_of_day": 3,
            "vehicle_type": 2,
            "traffic_density": 1
        }
    ]
}

📊 Model Performance

Metrics

• Accuracy: 89.2%
• Precision: 87.5% (weighted average)
• Recall: 89.2% (weighted average)
• F1-Score: 88.3% (weighted average)

Feature Importance

Feature	Importance	Impact
Speed	35.2%	Primary risk factor
Vehicle Type	18.7%	Vulnerability level
Weather	16.3%	Visibility/road conditions
Traffic Density	12.8%	Collision probability
Road Type	10.5%	Infrastructure safety
Time of Day	6.5%	Visibility/alertness

📈 Future Enhancements

Planned Features

• Real-time Data Integration - Traffic and weather APIs
• Geographic Factors - Location-based risk assessment
• Historical Analysis - Trend analysis and reporting
• Mobile App - Native iOS/Android applications
• Advanced Models - Deep learning implementations

Model Improvements

• Ensemble Methods - Combining multiple algorithms
• Feature Engineering - Additional risk factors
• Real-time Learning - Continuous model updates
• Explainable AI - Model interpretability features

Contribution Areas

• Model Optimization - Improve prediction accuracy
• Feature Engineering - Add new input parameters
• UI/UX Enhancement - Better user interface
• API Expansion - Additional endpoints
• Documentation - Code and usage documentation

👥 Authors & Contributors

• ARONAGENT - Project Creator & Maintainer

🙏 Acknowledgments

• Scikit-Learn Team - For the excellent ML library
• Django Community - For the robust web framework
• Traffic Safety Organizations - For research and data insights
• Open Source Community - For continuous support and inspiration

📞 Support & Contact

Getting Help

• 📖 Documentation: Check this README first
• 🐛 Bug Reports: Create an Issue
• 💡 Feature Requests: Suggest Features
• 💬 Questions: Start a Discussion

Connect with the Developer

• GitHub: @ARONAGENT
• LinkedIn: Connect for professional inquiries

🎯 Learning Outcomes

This project demonstrates:

• Machine Learning Integration with web frameworks
• Classification Model development and deployment
• REST API design and implementation
• Real-world Problem Solving with ML
• Full-stack Development with Django
• Model Deployment and production considerations

---

⭐ If you find this project helpful, please give it a star! ⭐

Built with ❤️ using Django, Scikit-Learn, and modern ML practices