Multi System Road Infrastructure Graph Dataset

The dataset is designed to enable controlled evaluation of scalability and complexity in graph-based Digital Twin (DT) frameworks. It consists of a series of synthetic graph instances of different sizes. These subsystems reflect realistic infrastructure domains and exhibit heterogeneous schemas and data characteristics: Spatial subsystem (InfraNet):Models spatial relationships using a reference element with associated bridges, road sections, and traffic counting points, including horizontal and vertical positioning. Bridge subsystem (Bridge):Represents hierarchical bridge structures with components, damages, inspections, and temporally repeated damage records. Road subsystem (Road):Captures road sections and lane-based components with periodic condition assessments. Traffic subsystem (Traffic):Contains time-series traffic recordings linked to counting points. The dataset explicitly incorporates: Schema heterogeneity across subsystems, Temporal heterogeneity through varying inspection and recording cycles, Distributed storage structure, where each subsystem graph can be deployed independently. To evaluate system scalability, dataset instances vary in size from 1 000 to 80 000 replicated base structures. In addition, alternative configurations are provided to assess system complexity, including: Reduced subsystem sets in a monolithic representation, Monolithic graph representations combining all subsystems into a single schema. Overall, the dataset enables systematic benchmarking of: Distributed vs. monolithic graph architectures, Scalability with respect to system size, Sensitivity to structural complexity and heterogeneity, Performance of distributed graph pattern matching.