Efficient and Scalable Semantic Data Ingestion for Smart City Digital Twin Platforms

Smart City Digital Twins are becoming fundamental tools to support decision-makers in facing challenges spanning in multiple domains of the urban environment. By mirroring the real city into a digital counterpart, and thanks to data and visual analytics, digital twins allow operators to perform analysis, predictions, and simulations to better plan the urban development. To achieve such goal, ingest and model semantically enriched Big Data is the first building block required to develop a digital twin and obtain a faithful replica of the city. Ontologies play a fundamental role in this regard; however, graph databases exploiting ontologies to set up knowledge bases are inefficient in handling big data time series and multiple formats. NoSQL solutions excel in storing big data. In this paper, a novel semantic driven process is proposed capable to store, enrich, and forward data messages providing support for semantic data retrieval taking into account relational, temporal, and geographic/spatial queries and real time event driven data. The solution is at the core of the open source Snap4City Digital Twin platform which is currently in use in the context of the Italian National Center for Sustainable Mobility, and for the national center on HPC in Italy.