Accepted in: Future Generation Computer Systems, 2020
The increasing availability of rating datasets (i.e., datasets containing user evaluations on items such as products and services) constitutes a new opportunity in various applications ranging from behavioral analytics to recommendations. In this paper, we describe the design of VugA, a visual enabler for the exploration of rating data and user groups. VugA helps analysts, be they novice analysts or domain experts, acquire an understanding of their data through a seamless integration between exploring users and exploring their collective behavior via group analysis. VugA is data-driven and does not require analysts to know the value distributions in their data. While automated systems can identify and suggest potentially interesting groups, they can do that for well-specified needs (e.g.,through SQL QUERIES or constrained mining). VugA helps analysts filter and refine their exploration as they discover what lies in the data. VugA enables analysts to easily acquire statistics about their data, form groups, and find similar and dissimilar groups. While most visual analytics systems are data-dependent, VugA relies on a data model that captures user data in such a way that a variety of group formation and exploration approaches can be used. We describe the architecture of VugA and illustrate its use via tasks and a user study. We conclude with a discussion on future work enabled by VugA.