Workflow Sharing Based on Semantics - A Case Study in Marine Spatial Planning
Summary
Marine Spatial Planning aims to strategically allocate human activities in marine areas, ensuring natural resources, ecological systems, and cultural heritage preservation. In order to achieve this objective, Marine Spatial Planning relies on an extensive collection of geographical data and simulation models. Preparing and analysing this myriad of data for the Marine Spatial Planning process could benefit from automated processes such as workflows. Nevertheless, issues with findability and reusability, such as incomplete metadata, lack of or incomplete provenance tracking of data products, and interoperability (not only synthetical but also semantical), have hindered this development. Therefore, this study uses web semantics and linked data principles to define abstract workflows to enhance findability, reusability, and track provenance. The workflow description includes the workflow metadata and describes the input and output data from each workflow step.
The abstract workflows are described using the Resource Description Framework (RDF) in Turtle format (TTL - Terse RDF Triple Language) since it is easily read by humans and machines, is supported by modern RDF libraries, is faster to read and avoids redundancy, being able to handle more complex graphs.
The proposed workflow description is used to define several workflows used in Marine Spatial Planning and stored in a triplestore; the choice for this database was GraphDB. Triplestore databases need to be queried using SPARQL; few users would know how to use this query language. Therefore, Sparnatural enabled users to create SPARQL queries through a graphic interface, and a webpage with the workflow diagram and metadata was created for each workflow.
In the end, a semantic search engine for a Marine Spatial Planning workflow repository was designed, and a functional prototype was developed with three components: a triplestore containing the workflows described, the SPARQL graphical query builder and the workflow webpages with the workflow diagram and metadata.
The thesis contributes to Marine Spatial Planning (MSP) by introducing a novel semantic web workflow definition that uses linked data principles. The new semantic web workflow definition improves the discoverability of workflows by integrating rich metadata and standardized vocabularies, enabling users to easily locate specific workflows relevant to their needs via semantic searches.
The workflow, workflow purpose and domain ontologies, helped achieve the research goals. It would be great to be able to support the user community in expanding the vocabulary to fit their needs as it evolves.
Future research is needed with a bigger sample of workflows and testers, and a few suggestions are given to improve the user experience. Furthermore, implementing systems that can translate concrete workflows from popular GIS into abstract workflows and vice-versa is in dire need.