Storm Guard: A Machine Learning Model for Enhanced Cyclone Prediction and Understanding

Traditional cyclone prediction systems rely on historical data and manual analysis, often lacking real-time accuracy and interactive visualization.To address these limitations, we propose an Advanced Cyclone Size Predictor that integrates four machine learning models (Random Forest, Linear Regression, XGBoost, and Neural Networks) with a Streamlit-based web interface for real-time predictions and comparative analysis.The system processes meteorological data (latitude, longitude, pressure, wind speed) to predict cyclone sizes (SiR34) and visualizes results on an interactive map.Key features include: Multi-model performance comparison (MAE, R metrics)Real-time geographical visualization (Folium/Plotly) Feature importance analysis Error prediction and trend visualizationThis implementation enhances disaster preparedness by providing data-driven, accurate, and explainable predictions for meteorologists and researchers.