Mod2Dash: A Framework for Model-Driven Dashboards Generation
Abstract
The construction of an interactive dashboard involves deciding on what information to present and how to display it and implementing those design decisions to create an operational dashboard. Traditionally, a dashboard's design is implied in the deployed dashboard rather than captured explicitly as a digital artifact, preventing it from being backed up, version-controlled, and shared. Moreover, practitioners have to implement this implicit design manually by coding or configuring it on a dashboard platform. This paper proposes Mod2Dash, a software framework that enables practitioners to capture their dashboard designs as models and generate operational dashboards automatically from these models. The framework also provides a GUI-driven customization approach for practitioners to fine-tune the auto-generated dashboards and update their models. With these abilities, Mod2Dash enables practitioners to rapidly prototype and deploy dashboards for both operational and research purposes. We evaluated the framework's effectiveness in a case study on cyber security visualization for decision support. A proof-of-concept of Mod2Dash was employed to model and reconstruct 31 diverse real-world cyber security dashboards. A human-assisted comparison between the Mod2Dash-generated dashboards and the baseline dashboards shows a close matching, indicating the framework's effectiveness for real-world scenarios.
Cite
@article{arxiv.2205.07204,
title = {Mod2Dash: A Framework for Model-Driven Dashboards Generation},
author = {Liuyue Jiang and Nguyen Khoi Tran and M. Ali Babar},
journal= {arXiv preprint arXiv:2205.07204},
year = {2022}
}