Predicting Road Accident Risk using LightGBM

The goal of this project is to predict a continuous accident_risk score for road segments based on a rich dataset of road, environmental, and temporal features. This is a classic regression problem with real-world implications for improving road safety. This project is structured into two main parts: 1-Exploratory Data Analysis (EDA): A comprehensive investigation of the dataset to uncover patterns, validate data quality, and identify key predictive features. 2-Predictive Modeling: The implementation of a robust machine learning pipeline to preprocess the data and train a final, optimized LightGBM model.