Use cases
SolarBench supports a range of research applications. We demonstrate three representative use cases that highlight the potential of its harmonized multimodal database and Python toolbox: standardized cross-site benchmarking, solar predictability analysis, and adaptation of irradiance nowcasting models to new PV systems (transfer learning). These use cases reveal insights that are difficult to obtain from fragmented datasets and inconsistent evaluation protocols. For details, please refer to our SolarBench paper.
Looking ahead, we hope to work with the community to explore new use cases, develop innovative methods, and contribute datasets, models, and evaluation tools that expand SolarBench’s capabilities and impact.