User Interaction Patterns for Linked Data

Tracking #: 3176-4390

This paper is currently under review
Authors: 
Mariana Aguiar
Sergio Nunes
Bruno Giesteira

Responsible editor: 
Guest Editors Interactive SW 2022

Submission type: 
Full Paper
Abstract: 
Linked Data is often still perceived as data that will only be consumed by machines, and not by humans as well. As a result, Linked Data applications still often use more traditional visualisations that come with usability issues. However, alternative user interaction approaches have been developed and evaluated, many of which have proven to be better solutions for inexperienced users. One technique to formalise and document these user interaction techniques and best practices is in the form of patterns. Here, we propose a pattern collection of 20 novel user interaction patterns for Linked Data resulting from the abstraction of common problems and solutions for visualising, searching, browsing, and authoring Linked Data. The proposed patterns are the combined result of the most common problems reported by members of the community, the knowledge and experience gathered in 10 pattern mining interviews, and the recurrent approaches collected through a literature review of the solutions for user interaction with Linked Data. To validate the structure of the pattern collection, we followed the Pattern Classification method, and to evaluate the adoption and quality of each proposed pattern, we conducted a pattern adoption survey. From this evaluation, we obtained positive scores for 18 out of the 20 patterns. We believe that the proposed pattern collection can be a valid and helpful tool for developers, regardless of experience, to improve the user interfaces of their Linked Data applications.
Full PDF Version: 
Tags: 
Under Review