Multi-scale urban energy data integration for energy poverty assessment: A semantic approach with evidence from the Netherlands

Urban energy poverty is a multi‑dimensional challenge, partly due to heterogeneous urban energy data across multiple spatial scales and domains. Existing assessment approaches often rely on isolated datasets or single-scale analyses, limiting their ability to support integrated and transparent evaluation of energy poverty. This study proposes a semantic approach for integrating heterogeneous urban energy data to support transparent and multi‑definition energy poverty assessment. It develops an ontology-based data integration framework, the Neighbourhood Energy Ontology (NEO), to connect building-level energy performance information, neighbourhood-level energy use, and aggregated energy consumption statistics within a semantic structure, thereby addressing the lack of interoperability. Moreover, a Random Forest model was applied to predict missing energy label data, demonstrating the synergy between machine learning techniques can be integrated with NEO and its associated dashboard, NEO Dash. The proposed approach is demonstrated through empirical evidence from a case study in the Netherlands. The results show that a semantic, multi‑scale integration approach improves the consistency of energy poverty assessment. The study highlights the analytical advantages of semantic technologies for urban energy research and discusses their implications for evidence-based analysis.