Facilitators: Kelly Gaither (UT-Austin) & TBD (LANL)

The visual cortex makes up roughly 33% by volume of the human brain. In contrast, hearing and taste make up approximately 3% and 8% respectively. In short, barring any visual impairment, our vision draws on the highest bandwidth channel to our brain — making it the fastest path to consume and synthesize data. It is this reason that visualization, or the process of making images out of data, is such a powerful tool for understanding complex systems. But, visualization in the absence of our ability to iterate with or interact, is of little value as the complexity and scale of problems increases.  If we marry interactive visualization with analytics rooted in data science principles, we create a feedback loop for query, analysis, and investigation that draws upon the strengths of each, and we refer to this process as Visual Analytics. With ever more challenging problems facing our world that require multidisciplinary expertise, advanced computing at scale, and the ability to iteratively reason, it seems only natural that we would address emerging challenges in Computational Visualization & Data Science. We focus on two different areas related to this broad theme:

  1. Human Data Interaction: The complexities of data have and will continue to be a focus of research. It has long been understood that a large percentage of time spent visualizing data is spent managing data. As the complexity of data increases: (heterogeneous, unstructured, multivariate, multiscale, time-varying, high-dimensional, etc.), the longer the time spent finessing this data into a consumable form. Perhaps it is time to rethink our approach to managing and understanding data. There is longstanding research in the area of human computer interaction that attempts to understand pain points and design around the fact that technology is a mediator between human and data. However, a more ideal solution would be one that facilitates human data interaction, enabling humans to interact directly with data rather than relying on technology as a broker. We can and should discuss models for this interaction that more closely model human-human interactions with established rules and protocols for interaction, design mechanisms to broker rapid response and rapid reasoning, and brokers for conflict resolution.
  2. Real-Time Decision Making: Decision-making is one of the cornerstones of visual analytics and has significant potential for advancing grand challenges in health and science. There has been significant research in decision-making in computational finance. However, there is much to be solved when there is a need to conduct real-time decision making in complex, dynamic scenarios. The ability to iteratively reason through large and potentially diverse and complex data streams has enormous challenges that increase exponentially as the scale of the problem increases. Augmenting this with informative visuals adds even more complexity. However, focusing on real-time decision making for complex global challenges has enormous potential for finding solutions to our society’s most pressing problems and for driving research in computational visualization and data analytics forward.