Editorial of transport data on the web

Abstract

Whether you are planning your next trip abroad or want a package delivered to your doorstep, chances are high that you will need a chain of services provided by multiple companies. Transport is inherently a geographically and administratively decentralized domain composed of a diverse set of actors, – from public transport authorities to vehicle sharing companies, infrastructure managers in different sectors (road, rail, etc.), transport operators, retailers, and distributors. As a result, it suffers vast data heterogeneity, which, in turn, brings severe challenges to data interoperability. However, such challenges have also been posed in other domains such as the Internet of Things [18], agriculture [11], building data management [17], biology [7] or open data [2], which have found their solutions using semantic web technologies. However, despite several research contributions [6,14,19,23,25], public-funded projects1,2 or academic-industry events,3,4 we have not yet seen a wide adoption of semantic technologies in the transport domain. We may only guess the inhibitors for adopting Linked Data in this domain: i) the SPARQL query language is not built for optimal path planning, and ii) RDF is perceived as highly conceptual by industry experts. We argue that SPARQL does not fit well with the concerns that typically matter to route planners (e.g., calculating the optimal Pareto path [4]). While calculating a path with SPARQL is feasible through property paths, controlling the path planning algorithm, which can hardly be done in SPARQL, is the core concern of route planners. On the other hand, the transport domain is dominated by different standards (e.g., NeTEx,5 or DATEX II6) and vocabularies, which are based on legacy data exchange technologies (e.g., XML or RDB). However, to construct a distributed and scalable architecture that addresses the current needs of this domain, the Web and its associated technologies (i.e., the Semantic Web) are the key resource.