🚗 Drivio – Intelligent Car Valuation Engine

Drivio is an AI-powered used car valuation system that combines machine learning, explainable AI (SHAP), business-aware pricing logic, uncertainty estimation, and negotiation intelligence into a single interactive web application.

It is designed as a complete AI product — not just a model.

---

🌟 Key Features

💰 Machine Learning Price Prediction

• Trained regression model for used car valuation
• Encodes categorical + numerical features
• Real-time prediction via Streamlit UI

📊 Explainable AI (SHAP Integration)

• Feature contribution breakdown
• Positive vs negative price impact
• Transparent reasoning behind predictions

🧠 Business-Aware Ownership Adjustment

• Single Owner → Premium adjustment
• Second Owner → Moderate depreciation
• Multiple Owners → Higher resale risk discount
• Clear separation between ML prediction and business logic

📈 Uncertainty Range Estimation

• Displays price confidence interval
• Reflects real-world market variability
• Improves trust in predictions

🤝 Negotiation Intelligence

• Buyer strategy suggestions
• Seller pricing guidance
• Risk-aware negotiation insights

🎨 Enterprise-Level UI

• Gradient animated brand header
• Interactive SHAP feature charts
• Vehicle preview by brand & model
• Clean valuation breakdown cards
• Startup-style polished design

---

🏗️ System Architecture

Streamlit UI
↓
CarPriceAgent (Orchestrator)
↓
Predictor (ML Model)
↓
SHAP Explainability
↓
Ownership Adjustment Layer
↓
Negotiation Agent

---

📦 Project Structure

drivio/
│
├── app/
│ └── streamlit_app.py
│
├── src/
│ ├── model/
│ │ └── predictor.py
│ │
│ ├── agents/
│ │ ├── price_agent.py
│ │ └── negotiation_agent.py
│ │
│ └── explainability/
│ └── shap_agent.py
│
├── artifacts/
│ ├── xgb_car_price_model.pkl
│ └── model_features.pkl
│
└── requirements.txt

---

🧮 Input Features

• Brand
• Model
• Fuel Type
• Transmission
• Body Type
• City
• Number of Owners
• Car Age
• Log(KM Driven)

---

🛠️ Tech Stack

• Python 3.12
• Scikit-learn
• XGBoost
• SHAP
• Pandas
• Altair
• Streamlit
• Joblib

---

🚀 How To Run Locally

1️⃣ Install Dependencies

pip install -r requirements.txt

2️⃣ Run Streamlit App

streamlit run app/streamlit_app.py

---

📊 Output Includes

• Final Estimated Price
• Estimated Market Range (Uncertainty Interval)
• SHAP Feature Impact Visualization
• Ownership Adjustment Breakdown
• AI Generated Market Summary
• Buyer & Seller Negotiation Strategy

---

🎯 What This Project Demonstrates

✔ Applied Machine Learning
✔ Explainable AI (XAI)
✔ Business Logic Integration
✔ Modular AI System Design
✔ Product Thinking
✔ Interactive Deployment

This project showcases end-to-end AI product engineering — from model training to explainable deployment.

---

🎯 Output Screenshot

1)User Input Interface

2)Valuation Interface

---

👨‍💻 Author

Aman Sharma
Focused on building intelligent, explainable AI systems.