MUREQ: a multilayer framework for analyzing and operationalizing visualization requirements

Understanding and interpreting vast amounts of information is pivotal in the contemporary data-rich age. Data visualization has emerged as a significant measure of comprehending these data. Similarly, an appropriate visualization can also enhance software modeling by providing straightforward and interactive representations. However, current data visualization methods predominantly require users to have data visualization-related expertise, which is usually challenging to obtain in reality. It is essential to bridge the gap between visualization requirements and visualization solutions for non-expert users, assisting them in automatically operationalizing their visualization requirements. This paper proposes a MUltilayer framework for analyzing and operationalizing visualization REQuirements that automatically derives appropriate visualization solutions based on users’ requirements. Specifically, we systematically investigate the connections among visualization requirements, visual variable characteristics, visual variable attributes, and visualization solutions, based on which we establish a conceptual framework that characterizes the relationships among different layers. Our proposal contributes to not only automatically operationalizing visualization requirements but also providing meaningful explanations for the derived visualization solutions. To promote our proposal and pragmatically benefit real users, we have developed and deployed a prototype tool based on the proposed framework, which is publicly available at https://reqdv.vmasks.fun. To evaluate our proposed framework, we conducted an initial controlled experiment with 44 participants to test the performance of the evolved mappings within our framework. Based on the expert’s feedback, we refined the mappings and incorporated a ranking system for visualization solutions tailored to specific requirements. To assess the current method, a subsequent experiment with another group of 44 participants and a focused case study involving two new participants were carried out. The results demonstrate that users perceive that the current method accelerates task completion, especially for complex tasks, by efficiently narrowing down options and prioritizing them. This approach is particularly advantageous for users with limited data visualization experience. Besides, the multilayer framework can be used to inspire the visualization of models in the software modeling community.