SOSA-SHACL: Shapes Constraint for the Sensor, Observation, Sample, and Actuator Ontology

The explosive growth of the Linked Data on the Web has greatly facilitated collecting data from remote sensors, from air quality sensors spread out across a city, to seismograph stations spread across the entire world. Integrating these heterogeneous data can be quite challenging; however one can achieve this through the use of available W3C standards to create a knowledge graph. For this use case, the W3C also provides a standard, the Sensor, Observation, Sample, Actuator (SOSA) Ontology, that allows for the semantic encoding of sensors and their observations. However, even with the guidance of this standard, it may be difficult to produce a correct graph with high fidelity from heterogeneous sources. In this paper we present a set of (data) shape constraints, called SOSA-SHACL, for the SOSA ontology using a data validation language, namely the W3C standard SHACL (Shape Constraint Language). These constraints enable us to evaluate whether the modeled observations in our Knowledge Graph comply with the SOSA recommendations. Furthermore, we show through several case studies how the closed world assumption plays a role in the process of designing such shape constraints, especially as SOSA is based on the open world assumption.