Context
Rooftop solar projects often require fast, remote assessment of building suitability and expected generation potential to accelerate installations and prioritise sites.
Challenge
Scaling rooftop solar deployment requires a repeatable way to assess rooftop characteristics (e.g., usable area, shading, and solar potential) across many buildings without relying solely on time‑intensive, manual site assessments.
AI Solution
The Google Maps Platform Solar API generates detailed rooftop data using Google’s geospatial data and computing resources to evaluate rooftop solar energy potential.
It provides three main endpoints: buildingInsights (location, dimensions, solar potential), dataLayers (raw solar datasets for an area), and geoTiff (rasters such as a digital surface model, imagery, annual/monthly flux maps, and hourly shade).
Impact & evidence
Solar API enables early‑stage solar feasibility analysis to be carried out remotely and at scale, supporting faster screening of buildings and more consistent system design inputs. Benefits include efficiency gains in solar assessments, improved quality of proposal development, and earlier availability of data to support customer discussions.
Implementation / Adoption
Adoption requires configuring a Google Cloud project and calling the API endpoints to retrieve building insights and solar data layers for targeted locations.
Coverage varies by country/region and is provided by Google via a country‑by‑country coverage listing.
Data
Solar API is built on Google‑generated geospatial datasets, including aerial and satellite imagery, digital surface models (DSMs), rooftop characteristics, and modelled solar irradiance and shading at hourly, monthly, and annual resolutions. Both processed insights and underlying raster datasets are accessible via Data Layers and GeoTIFF endpoints. The solution does not rely on customer‑provided grid, metering, or operational data.
Future enhancements
The Solar API is launching an experimental feature that expands coverage to previously unsupported regions, based on machine learning models applied to satellite imagery.