Solar Energy Forecasting in Seattle Using Machine Learning Models

Abstract

Seattle faces challenging environmental issues, which encourage people to explore ways of improving the utility of renewable resources, such as Global Horizontal Irradiance (GHI). This research consists of methods of data cleaning, data visualization, and designs of Machine Learning models, like Linear Regression, Decision Tree, and Random Forest. Based on the results of R2 and RMSE values, the Random Forest has the best performance among other models, with an RMSE≈of 48.1 and an R²≈of 0.82. The research is dedicated to providing a stable and reliable prediction of GHI Values. With GHI predictions, city planners and government are able to efficiently plan and manage solar resources, mitigate instability risks, and enhance building energy performance through improved control of lighting and shading systems. Moreover, the application of machine learning models is a good start in environmental sciences and gives a solid foundation in the application of data sciences in real life. Additionally, the findings offer a reference framework for integrating predictive solar energy data into urban planning and renewable energy policy-making. Future work could extend these models to incorporate real-time data for dynamic forecasting and decision support.