You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Townsend snow loss (via pvlib) + precipitation-aware soiling model using monthly climate inputs (manual-clean option); month-by-month results to support EPC yield estimates and O&M cleaning plans.
Deployment and a public backend URL will be available soon, as I am currently limited by hosting costs on services such as Railway, Docker, Kubernetes, and AWS
Most solar forecasters throw raw ML at the weather. This one doesn't - a clear-sky physics model + plane-of-array transposition set the baseline, and ML learns only the residual. Beats clear-sky and persistence (0.85 / 0.69 skill) on real data. Python, scikit-learn, live Streamlit demo.