| dc.description.abstract | This thesis presents a visual representation approach for Linked Data mappings known as Jigsaw Puzzles for Representing Mappings, or Juma. The term Linked Data refers to a set of best practices for publishing and interlinking data on the Web. A Linked Data dataset is structured information encoded using the Resource Description Framework (RDF), in which resources are identified by and linked with other datasets using HTTP URIs. Linked Data datasets cover a wide range of knowledge domains, where often concepts overlap. In such cases, mappings can be created to reduce heterogeneity and facilitate the consumption of information by informing agents which concepts are related, and how. These types of mappings are called semantic mappings. Another area in which we find use for mappings is when transforming data from one representation to another ‘ from non-RDF to RDF for example. We call those mappings uplift mappings. Producing such mappings can be difficult, even for experts in Semantic Web technologies, requiring knowledge on the specifics of the mapping language being used as well as significant amount of human effort for their creation, modification, curation and maintenance. Nonetheless, literature suggests that this user involvement is fundamental for producing quality mappings. Suitable visual representations may be used to support user involvement and alleviate the knowledge required for producing Linked Data mappings. Through a systematic literature review, a set of requirements for a visual representation for Linked Data mappings were defined. Juma was then proposed as a novel approach, based on the block metaphor, for the representation of mappings in Linked Data. The block ‘ or jigsaw ‘ metaphor was chosen as it takes advantage of the user’s familiarity to jigsaw puzzles, fosters users to explore the combinations of blocks, and for being accessible to experts and non-experts alike. Juma leverages the use of the block metaphor in order to facilitate the interpretation of mappings in Linked Data. In Juma, blocks are used to abstract and capture different mapping constructs, where the connection of the different blocks form a mapping. Each block is then translated to an equivalent mapping representation, which can be done for distinct mapping languages. The Juma approach was evaluated through five experiments categorized in three aspects: creation (and editing), understanding, and expressiveness. The creation of mappings was evaluated through two user experiments, where participants were asked to create a mapping using applications that apply the Juma approach for the representation of mappings. Another user experiment was conducted to evaluate the understanding of mappings represented using a Juma application. Finally, two experiments were conducted to evaluate the expressiveness of the Juma approach in the representation of uplift and semantic mappings, respectively. These evaluations indicated that the Juma approach is effective in representing uplift and semantic mappings, and that it aids users in the creation, editing and understanding of Linked Data mappings. The research in this thesis has yielded one major contribution and three minor contributions. The major contribution is the design and development of the Jigsaw Puzzles for Representing Mappings (Juma) approach. The first minor contribution is the Juma R2RML application. Juma R2RML applies the Juma approach to the R2RML mapping language. The second minor contribution is the Juma Uplift application. Juma Uplift has a higher level of abstraction in order to be able to generate mappings using multiple distinct mapping languages. The Juma R2RML and Juma Uplift applications apply the Juma approach in the representation of uplift mappings. The third minor contribution is the Juma Interlink application. Juma Interlink applies the Juma approach in the representation of semantic mappings. | en |
