## Overview

**Can you predict traffic accident risk from driving behavior?**

This competition challenges you to build models that classify driving risk levels using real-world time series data from urban driving sessions. Using sensor data from vehicles, drivers, and environmental conditions, you'll predict whether a driving situation poses low, medium, high, or very high accident risk.

### The Challenge

Modern vehicles and wearable devices generate rich streams of data about driving behavior, vehicle performance, and driver health. This competition explores whether machine learning can identify dangerous driving situations in real-time by analyzing patterns in:

- **Vehicle telemetry**: Speed, acceleration, engine performance
- **Driver health**: Heart rate monitoring
- **Environmental conditions**: Weather, visibility, precipitation
- **Road characteristics**: Historical accident data, design speed
- **Temporal patterns**: Time of day effects

### Why This Matters

Traffic accidents remain a leading cause of injury and death worldwide. Developing accurate risk detection systems could enable:

- Real-time driver alerts and assistance systems
- Insurance telematics and risk assessment
- Fleet safety monitoring
- Road infrastructure improvements
- Traffic safety research

### The Data

The dataset is derived from the **POLIDriving** dataset, collected from real driving sessions in Quito, Ecuador. It includes:

- **1,980 labeled training samples** (balanced across 4 risk levels)
- **23,152 test samples** for final evaluation
- **82,223 unlabeled samples** for semi-supervised learning (optional)
- **15 features** from vehicle sensors, health monitors, weather APIs, and traffic databases
- **18 hours** of real-world driving data

### Your Goal

Build a model that accurately classifies each driving observation into one of four risk levels:

1. **Low Risk** - Normal, safe driving conditions
2. **Medium Risk** - Elevated risk requiring attention
3. **High Risk** - Dangerous conditions requiring intervention
4. **Very High Risk** - Critical situations with imminent danger

Good luck, and drive safely! 🚗