Aerial image analysis at scale is a key tool in gathering this data, especially for tracking renewable energy installations like rooftop photovoltaic (PV) systems. Scalable methods for identifying PV installations can help map out the built environment, providing insights on energy resource availability and helping to optimize urban planning and infrastructure development. The project uses a ResNet- 34 neural network classifier trained to detect PV installations on rooftops from aerial images. By combining OpenStreetMap data with aerial imagery, building rooftop images are extracted, cropped, resized, and prepared for classification. Source code and results are available on the MODERATE GitHub repository (https://github.com/MODERATE-Project/building-stock-analysis.git) and MODERATE Zenodo community.