On May 19, 2012, a „Visualization Curriculum Workshop“ was held at the University of Cagliari, Sardinia, sponsored by the SIGGRAPH Education Committee. It was a follow-up on the Visualization Curriculum Panel at Eurographics 2012. Participants were David Ebert, Beatriz Sousa Santos, Holly Rushmeier and Gitta Domik.
During the morning, workshop participants followed up on the issues as discussed at the Eurographics Panel “Visualization Curriculum Panel – or the Changes We Have Made to Our Visualization Courses over the Last 10 Years”. Since the early 1990s, teaching Visualization has been part of the Computer Science curriculum at many universities world-wide, but applications, methods and technologies have changed since then. It seems currently appropriate to make a distinction between
- data visualization,
- information visualization and
- visual analytics
though many commonalities exist and all of these can be summarized under the heading of “visualization”. To shortly summarize some significant data differences,
- in data visualization, data dimensions usually coincide with physical dimensions, such as in medical or remote sensing or flow data,
- information visualization typically deals with multidimensional data as in finances, business intelligence, or large databases and mostly physical dimensions are not present
- in visual analytics, data is characterized by large, complex, and heterogeneous data sets.
The rest of the workshop was spent in discussing the following proposed outline of the following 11 themes to teach “visualization”, where additions and distinctions must be used if focusing on a specific type of visualization.
The 10 themes were taken from previous curriculum suggestions and are as follows:
- Definitions and History of Visualization.
- Data. This includes models, transformations, data characterization as well as data manipulations
- The User. This includes the human information processing limitations and capabilities as well as the task that a user brings to the visualization problems
- Design stage. This stage describes a careful mapping of data components to visual attributes
- Visual Presentation Techniques. This should include a wealth of visualization solutions, sorted by data characteristics, by application domain, and/or by task, and described by their various parameters. This theme can be presented at the breath-level by showing and discussing (interactive) visualizations. It can be trained by using available tools at the breadth-level or in-depth by developing interactive visualization techniques on a GPU, or in any breadth/depth stage in between, depending on the qualifications of students.
- Interaction techniques. While interaction techniques are a demand per se for Visual Analytics, they are increasingly also present and demanded for data and information visualization, where GPU techniques can reach the necessary processing speed.
- Communication. A focus on production, presentation and dissemination as part of the visualization process was raised by visual analytics, but needs to be observed as well for data visualization and information visualization.
- Collaboration. Interactivity aids collaboration, but other issues are of concern in the collaboration among stakeholders.
- Evaluation. Evaluation is seen as a continuous process, starting with the requirement analysis of the visualization problem, continuing as a constant awareness process of the human-in-the-loop of software processes towards a visualization goal, and ending with evaluation techniques to secure reaching this goal for the specific visualization problem.
- Displays. Different capabilities of displays (size, memory, processors) pose problems on visualization techniques, interactivity and communicaton and need to be treated at least at the Mapping/ Design stage.
It was decided to have these preliminary discussions continued at the BOF "Visualization and Visual Analytics Curriculum" at SIGGRAPH 2012.