MetaViz – A graphical meta-model instantiator for generating information dashboards and visualizations

Abstract

ObjectivesThis work presents the application of meta-modeling to the data visualization and dashboards' domain to obtain two main products: a system that allows the instantiation of the meta-model through a graphical interface and a code generator that takes the instantiation of the meta-model as an input to generate visualizations and dashboards. MethodsA domain engineering approach complemented with an example-driven methodology was iteratively employed to develop the dashboard meta-model. This meta-model was subsequently used as an input to implement a code generator of information dashboards. These two artifacts were finally combined to design and develop the architecture of MetaViz. ResultsThrough this process, it was possible to generate visualizations and dashboards using visual elements and basic interactions. MetaViz allows the generation of basic charts (line charts, scatter plots, pie charts, etc.) as well as more complex displays with interactive behavior along different views, layouts, and operations. ConclusionsThe development of MetaViz has served as proof of the viability and benefits of applying these methodologies to a complex domain, but also to set the foundations of a system that allows users to trace every single element from a data visualization to its most primitive values.