skg2k05/SolarPredict — reverse-engineered prompt

Build me a simple SolarPredict app that predicts how much solar power a plant will produce from weather conditions. Use the provided generation and weather sensor CSV data to train a Linear Regression model that predicts AC power in kW from irradiation, ambient temperature, and module temperature.

I want a clean Streamlit interface where a normal user can move sliders for those three inputs and instantly see the predicted solar output. Also show a short performance message or recommendation, like whether the conditions look good for generation or if output may be low. Include basic charts or metrics so the project feels complete, such as model accuracy, error values, and a simple comparison of actual versus predicted output.

Please keep it easy to run locally, save the trained model, and include a notebook or script that explains the training process. Make the UI friendly for schools, homeowners, or small industries who want to estimate solar generation. Look up current docs online if you need to.