A new frontier in dashboard design: Evaluating an innovative meta‐modelling approach through expert insights

• Published in: Expert systems
• Main Authors: Vázquez‐Ingelmo, Andrea, García‐Holgado, Alicia, García‐Peñalvo, Francisco José, Therón, Roberto, Colomo‐Palacios, Ricardo
• Format: Journal Article
• Language: English
• Published: 05.02.2024
• ISSN: 0266-4720; 1468-0394
• DOI: 10.1111/exsy.13560

Abstract Context Data visualizations and dashboards are essential in disseminating information to broad audiences. However, designing these instruments is complex due to the multitude of involved factors. Theoretical frameworks are therefore critical to guide the design and implementation of data visualizations. Objective This article presents the approach used to pinpoint key features that impact the creation of information visualizations, organize these features using meta‐modelling, and validate the final artefact through expert judgement. Method Meta‐modelling facilitates the distillation of characteristics and functionalities critical for the development of information systems. Utilizing this method, a dashboard meta‐model was constructed in preceding research, yielding a framework of abstract components and connections to delineate specific products. This meta‐model serves as a robust tool for the automated generation of information visualizations and dashboards. Nonetheless, to confirm its quality and efficacy, it underwent a thorough quality assessment and expert evaluation to identify and address potential shortcomings. Results The validation process resulted in positive ratings and constructive feedback to enhance the meta‐model's present iteration. Experts mainly highlighted the meta‐model's complexity and the clarity of certain entities, noting potential confusion for non‐experts. Conclusions While the meta‐model is confirmed to be a coherent and pertinent tool, it requires additional clarifications to simplify its complexity. Addressing this, graphical interfaces have been developed to enable users to interact with the dashboard meta‐model—and to create dashboard instances—through straightforward interactions.