Studying human analyst's behavior in automated tracing is a new research thrust. Building on a growing body of work in this area, we offer a novel approach to understanding requirements analyst's information seeking and gathering. We model analysts as predators in pursuit of prey - the relevant traceability information, and leverage the optimality models to characterize a rational decision process. The behavior of real analysts with that of the optimal information forager is then compared and contrasted. The results show that the analysts' information diets are much wider than the theory's predictions, and their residing in low-profitability information patches is much longer than the optimal residence time. These uncovered discrepancies not only offer concrete insights into the obstacles faced by analysts, but also lead to principled ways to increase practical tool support for overcoming the obstacles.

For many software projects, keeping requirements on track needs an effective and efficient path from data to decision. Visual analytics creates such a path that enables the human to extract insights by interacting with the relevant information. While various requirements visualization techniques exist, few have produced end-to-end values to practitioners. In this paper, we advance the literature on visual requirements analytics by characterizing its key components and relationships. This allows us to not only assess existing approaches, but also create tool enhancements in a principled manner. We evaluate our enhanced tool supports through a case study where massive, heterogeneous, and dynamic requirements are processed, visualized, and analyzed. In particular, our study illuminates how increased interactivity of requirements visualization could lead to actionable decisions.

Clustering is of great practical value in discovering natural groupings of large numbers of requirements artifacts. Clustering-based visualization has shown promise in supporting requirements tracing. In this paper, we transform the success to a wider range of clustering-based visual exploration tasks in requirements engineering. We describe ReCVisu, a requirements exploration tool based on quantitative visualizations. We discuss the key features of ReCVisu and its potential improvements over previous work.