AI-Driven Hurricane Forecasting Model

I developed a machine learning model that predicts hurricane classifications ahead of time using historical data from the NOAA National Hurricane Center’s HURDAT2 database, which contains records dating back to the 1800s. I wrote a custom Python preprocessing pipeline to parse and clean raw hurricane data, converting it into a structured dataset suitable for training.

Using a Random Forest classifier, the model learns patterns in storm behavior and achieves approximately 95% accuracy in identifying F1 hurricane classifications before they occur. Because the model is trained on NOAA’s freely available datasets, it offers a cost-effective and scalable AI solution for disaster preparedness, forecasting, and risk assessment.