Information extraction for constructing a Knowledge Graph of Positive Energy District

Abstract

Positive Energy Districts (PEDs) are discussed across a wide range of policy papers, research studies, and project reports. Yet the terminology and conceptual boundaries used to describe them remain fragmented and inconsistent. This study examines how ontology engineering and Knowledge Graph (KG) methods can bring structure and clarity to this diverse body of material by providing a transparent and reproducible representation of PED knowledge. A study-specific PED ontology was developed to capture recurring concepts in the literature and was aligned with SAREF4City to incorporate relevant smart‑city semantics. Using this ontology as a foundation, a KG was constructed from a curated corpus of PED documents by extracting sentences, generating semantic triples, and aligning them with the ontology. The resulting KG links documents, sentences, entities, and conceptual relations, enabling structured retrieval across technical, governance, process, financial, environmental, legal, managerial, and social dimensions. Querying the KG revealed recurring patterns, highlighted dominant themes, and showed that Governance, Process, and Technical concepts appear far more frequently than other aspects, indicating an imbalance in how PEDs are documented. While the approach depends on sentence‑level extraction quality and reflects uneven coverage in the underlying corpus, it offers a practical way to support PED planning and implementation. The framework can assist with tasks such as technology selection, governance design, and evidence‑based decision‑making, and contributes to more transparent, accessible, and socially meaningful PED development. By improving the clarity and accessibility of PED knowledge, the approach supports more inclusive and sustainable urban transformation.