Smart City Data Modelling using Semantic Web Technologies

In modern Smart City ecosystems, the wave of Big data has offered the potential for cities to obtain actionable insights from a large amount of data, collected from heterogeneous sources. To cope with data variety and to support querying facilities and navigation, domain experts, who possess the knowledge required to understand data sources structure and content, resort to the definition of metadata layers, for reconciling terminological differences and resolving ambiguity of data sources attributes. To this aim, the adoption of Semantic Web technologies assures a unified view of heterogeneous data sources and improves data access, leveraging open standards for knowledge representation. In this paper, we propose an approach to create a semantic data layer over Smart City data sources. The approach is grounded on two main pillars: (i) a procedure, to build and manage the semantic layer; (ii) a tool, implementing the steps of the procedure, devoted to support domain experts. Results of an experimentation conducted on representative Smart City data sources demonstrated the effectiveness of the approach. In particular, the presence in our approach of a preliminary step regarding lexical enrichment of data source attributes helped increasing the retrieval of useful concepts from domain ontologies for attribute annotation. In addition, allowing domain experts to provide additional ontological knowledge by specialising or creating new concepts and semantic relationships fostered flexibility in the annotation procedure, facilitating the evolution of the semantic layer. Finally, the developed semi-automatic tool helped domain experts in the semantic annotation process by also providing validation support.