VoxSolaris: Code, Data, and Results for LiDAR-Based Voxel Ray Tracing and Rooftop PV Shading Validation in Kuopio, Finland

This archive contains the code, input data, processed LiDAR products, notebooks, shadow attenuation matrices, and validation outputs supporting the VoxSolaris rooftop photovoltaic shading study in Kuopio, Finland. VoxSolaris is a Python-based workflow for modelling solar transmittance in complex peri-urban environments using airborne LiDAR-derived voxel scenes. The workflow represents buildings as opaque geometry and vegetation as semi-transparent canopy volumes using Beer-Lambert attenuation, then applies cone-based ray tracing to account for the finite angular diameter of the solar disc. The resulting directional Shadow Attenuation Matrices are used to correct baseline photovoltaic production estimates from the Finnish Meteorological Institute PV forecasting model. The archive includes scripts and notebooks for voxel-based shadow-matrix generation, Bank 1 and Bank 2 PV analyses, all-day evaluation, manuscript figure generation, daily diagnostic plots, 3D ray/scene visualizations, PV and meteorological input files, processed LAZ point-cloud files, validation metrics, and the packaged FMI PV forecasting dependency used in the analysis. The case study evaluates 5-minute rooftop PV inverter measurements and meteorological observations for the 2021 growing-season period, with specific emphasis on clear-sky shading behaviour and full-period PV yield prediction.