Crash report analysis

• Tech Stack: Python, Scikit learn

- This project analyzes New York City collision data to predict crash severity and identify influential factors such as vehicle make, driver behavior, and crash circumstances.

- Using machine learning techniques like Random Forest, XGBoost, and Neural Networks, it achieved high accuracy in predicting collision outcomes and revealed critical insights for targeted safety measures.

- The study integrates data preprocessing, feature engineering, and visualization to guide policy interventions under the Vision Zero initiative, aiming to enhance urban traffic safety.