21 January 2026 in IEEE Access
Building-specific knowledge, such as building type and function information, is important for numerous energy applications. However, comprehensive datasets containing this information for individual households are lacking in many regions of Europe. For the first time, we investigate whether predicting building types and functional classes at a European scale is feasible based only on open datasets available across countries. We train a graph neural network (GNN) classifier on a graph dataset of OpenStreetMap (OSM) buildings across the EU, Norway, Switzerland, and the UK. To efficiently perform training using the large-scale graph, we utilize localized subgraphs. A graph transformer model achieves a Cohen’s kappa coefficient of 0.7536 when classifying buildings into 9 classes, and 0.8442 when classifying buildings into the residential and non-residential classes. Transferability experiments demonstrate strong generalization to unseen regions ( $\kappa =0.7379$ in region-based spatial cross-validation), with a model trained in Germany maintaining predictive power across Europe when minimal local data is included ( $\kappa =0.6403$ ). The results imply three key novel contributions to the literature. Firstly, we show that building classification across multiple countries is possible using a multi-source dataset. Secondly, our results indicate that GNNs that consider contextual information about building neighborhoods improve predictive performance compared to models that only consider individual buildings and ignore the neighborhood. Third, we show that training with GNNs on localized subgraphs instead of standard GNNs improves the performance of the building classification task.