Abstract:
Human decision making by nature is predominantly a cognitive process conducted in an individual’s mind consciously or subconsciously to select an optimal solution from multiple alternatives for a certain decision goal based on pre-defined evaluation criteria. It is significantly influenced by the individual’s ability to acquire situation awareness and form mental models of the interested decision problem. With the rapid advances in the domain of information visualization, many visualization techniques and systems have been widely adopted to amplify situation awareness, and to inform, shape, change and/or reinforce mental models and behaviors of people. They do so by supporting the development of visualizations that fulfill particular purposes under certain contexts. Typical visualization purposes are to address issues within a diverse range of application domains, support user activities involved in information discovery and interpretation and decision making, and express thoughts/feelings/emotions on life, world, issues, etc. The effectiveness of fulfilling such purposes can be largely affected by both the decisional problem context within which the visualizations are deployed and the situational context of their stakeholders. A key feature different visualizations share in common is that visualization purposes, contexts and stakeholders may vary over time and domain. Such changing facets of visualizations often result in dynamic changing visualization requirements. Moreover, visualization problems nowadays often incorporate increasingly high complexities that may emerge from enormous data volume, complex data types, the need for integrating multiple visualizations, and/or multiple paradigms/domains involved. These issues significantly exacerbate the difficulties involved in the development of effective visualization solutions. Dealing with such complex problems often requires visualizations to be developed in a manner that they can be flexibly created, instantiated, manipulated, customized, integrated, executed and modified. Though many existing visualization techniques and systems tend to provide reasonable support for particular paradigms, domains and data types, they are weak when it comes to (1) addressing the complexities involved in the problems and (2) supporting the ever-changing purposes, contexts and stakeholders, in a manner that sustains visualization effectiveness. To address the above problems, issues and requirements, we suggest that visualizations be viewed from the intertwined triumvirate perspectives of purpose, context, and stakeholder. We introduce purposeful visualizations to encapsulate these concepts. We define purposeful visualizations as visualizations that fulfill a particular purpose for one or more stakeholders within a certain context. The concept of purposeful visualizations highlights the status of a visualization with regard to its visual capability of accommodating an intended purpose; it also accentuates the changing nature of visualizations caused by stakeholder, purpose, and context. Every visualization can become purposeful under a certain context but may fail to retain its purposefulness and effectiveness when the context is changed. The objectives of our research are to (1) formally define the concept of purposeful visualizations, (2) propose a model that facilitates understanding, creating and evaluating purposeful visualizations, (3) develop a process to guide the application of purposeful visualizations for addressing various visualization purposes in the context of decision making, (4) design purposeful visualization system frameworks and architectures to support the creation, instantiation, modification, execution, integration, transformation and adaptation of visualizations, (5) validate these concepts, models, processes, frameworks and architectures by implementing a prototypical purposeful visualization system that can be used to create purpose-driven, context-sensitive, stakeholder-relevant visualizations, and finally (6) demonstrate the functionalities and capabilities of the prototype through a sequence of real world scenario-driven illustrations drawn from the utility sector.