The growing deployment of rooftop photovoltaic (PV) systems in urban environments requires accurate, scalable, and reproducible tools for estimating solar radiation and PV energy production at fine spatial and temporal resolutions. Existing GIS-based solutions are often limited to yearly or monthly simulations and lack the modularity needed for large-scale, real-sky assessments. This thesis presents the design and implementation of a GIS-based backend platform for solar radiation simulation and PV production analysis in urban contexts. The proposed system is built around a multi-stage geospatial processing pipeline that integrates Actinia/GRASS GIS for raster-based terrain and radiation computations, and exposes all functionality through a REST API implemented with FastAPI.