Efficient storage and querying of complex building data are critical challenges in Building Information Modelling (BIM), particularly for architects in the early design and planning stages. This paper systematically compares two graph-based data models, Resource Description Framework (RDF) and Labeled Property Graphs (LPG), using BIM datasets, in order to assess their suitability for supporting design exploration and alternative evaluations. Benchmark tests of three graph databases,Neo4j, GraphDB, and Blazegraph, analyze loading times, storage requirements and query performance. The results indicate that LPG-based systems support more efficient querying of explicit design information, while RDF-based systems provide stronger support for semantic modelling and reasoning. Herein, the choice of graph representation and the impact of query languages significantly influences the effectiveness of graph-based decision-support in early design stages. These insights guide software developers in selecting graph technologies to improve architects' and engineers' early design decision-making, while additionally highlighting opportunities for integrating AI-driven applications. • Empirical comparison of RDF and LPG graph representations for BIM decision-support. • Benchmark of Neo4j, GraphDB and Blazegraph using identical BIM datasets and queries. • Analysis of query performance, loading time and modelling implications in early design. • Evaluation of graph representations for variant exploration and element identification. • Assessment of RDF and LPG suitability within CBR-based building design processes.
Napps et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: