This study focuses on Spatio-Temporal Modelling for Cyclone Track Forecasting, leveraging Machine Learning (ML) and Deep Learning (DL) techniques to predict the direction of cyclones based on historical and real-time data. Accurate cyclone tracking is crucial for disaster preparedness, and this model integrates key meteorological variables such as latitude, longitude, wind speed, wind direction, pressure, temperature, humidity, and distance from land to determine the cyclone's movement. The system uses a combination of LSTM (Long Short-Term Memory), ANN (Artificial Neural Networks), Random Forest, and XGBoost algorithms to predict the cyclone's direction—whether it moves towards the East, South, West, or North. The developed model provides an efficient way to forecast cyclone tracks, thus aiding authorities in timely evacuation and safety measures. The solution utilizes a web-based application built with HTML, CSS, Bootstrap, JavaScript, and backed by Python with Django for web framework support. The data processing and model training are powered by Pandas, NumPy, Scikit-Learn, and TensorFlow, with SQLite used for database management.

Keywords: Spatio-Temporal Modelling, Cyclone Track Forecasting, Machine Learning, Deep Learning, LSTM, ANN, Random Forest, XGBoost, Cyclone Direction Prediction, Meteorological Data, Web-based Application, Python, Django, TensorFlow, SQLite.