Contributors: Sherry Hsi and Judi Fusco
Scientific visualization has had a tremendous impact on STEM professions and the practice of science over the last decade by capitalizing on the power of the human visual perception system to identify patterns in complex data. Visualizations can range from simple drawings; to interactive computer-based animations; to 3D models formed from assembling multiple massive data sets; to colorful simulations of scientific phenomena enabling the exploration and demonstration of relationships between and among different variables. These visualizations can be rendered from real-time data, manipulated by end users, shared, and linked to other representations. They can be used to engage learners in multiple settings including classrooms, homes, and informal science institutions. There is a rich and long history of research on visualizations and how they can aid the cognitive processes involved in learning (although mostly in laboratory and some classroom settings).
The design of effective data visualizations for learning has been hampered by several factors related to the availability of advanced technologies and design tools for visualization, as well as the limited number of visualization designers with backgrounds and cross-training in STEM, media arts, graphic design, computer sciences, and learning sciences. With available complex data sets, faster graphics capabilities, and new cyber-infrastructures, we are now capable of supporting more interactive and realistic virtual worlds, phenomena, and 2D and 3D imagery. Moreover, new collaborations are forming around STEM professionals, learning scientists, media artists, and evaluation professionals who can collaboratively generate new design knowledge for producing and effectively assessing learning from visualizations. How to advance learning and teaching through better design and integration of visualizations into new learning environments continues to be a critically important area of research.
Transformative Potential
Cyberinfrastructures are enabling real time data to be collected from networked seafaring research vessels, space probes, global sensor networks, remote satellites, public databases, and citizen scientists. Many of these data sources are open and available for education and public use. Using data sources such as these can be exciting for students and teachers. With more fundamental understandings of how people learn, combined with digital tools and techniques that were originally developed for professionals, visualization environments can be better used to support learning and teaching.
Visualizations have the power to allow students to see phenomena and processes that are too tiny (e.g., nanomanufactuing), too fast (e.g., chemical reactions), too large (e.g., super novas), or too complex (e.g., protein folding) to be seen by the naked eye. Informed by research on visual learning and design, a large-scale dataset can be turned into an inquiry experience and virtual experiment where students and teachers together manipulate and transform a visualization, sparking new hypotheses and questions to explore by creating further visualizations to enhance understanding. They can also be used to compress long time scales to create elegant useable representations for the short period of classroom instruction.
Importantly, visualizations can be shared. As expressive products, they support the communication of ideas between the designers and the consumers of the visualization; though, perhaps even more powerfully, they provide a rich shared context for exploration and discourse among learners sitting next to each other…or across the world.
Vision
In 10 years…
On the cyber side, Hardware platforms and accompanying software for color 3D visualizations will be sufficiently low-cost that they will find their ways into schools, homes, and afters-chool centers. 3D glasses will be like gigabyte thumb drives today.
On the learning side, a body of design knowledge, expertise, and cognitive research will be available to inform the design of visualizations for learning (see Challenges for more on what we need to know). Schools will embrace teaching media arts and data visualization as part of a their regular curricula.
A rich set of software will be available via the Web that will enable students to create their own visualizations to show how they currently think about phenomena and understand conceptual models. (Concurrently, advanced assessment systems will be able to score these, provide feedback, and suggest alternative visualizations to create and/or view.) During lesson planning, teachers will have access to discipline-specific visualization tools that enable them to quickly and easily create a visualization including reflection prompts and alternate pathways through a visualization.
A learner who wants to “see” and “feel” how a mitochondria works can search an open digital library, locate the link to a haptics-friendly visualization, put on their active 3D glasses, and then explore and play inside of a virtual animal cell. Similarly, a family visiting a local museum can use an interactive learning visualization exhibit to change the magnitude of a simulated earthquake to see its impact on the height of a tsunami on their town.
Visualizations will be both more expressive and more constructive.
Challenges
One set of challenges are around design principles for visualizations. Massive data sets alone don’t generate good visualizations. Visualizations need to be thoughtfully designed and mediated to be useful for learning. We need a better understanding of what kinds of visualizations work and what kinds fail (and when). For example, how do we design them for difference audiences and settings, e.g., formal and informal? Facilitated vs. standalone use? For experts versus novices? Research shows that learners interpret visualizations differently based on their domain expertise and background knowledge. As we gain understanding around these questions, we can articulate the principles around the design of visualizations, for different kinds of audiences, and hopefully avoid failures. Design principles should also include guidance around how visualizations can be combined with narrative voice-overs, graphic elements, labels, and other designed media.
Another set of questions occur around how visualizations can be effectively deployed as teaching tools? For example, how should disciplinary knowledge be extracted and depicted in a teaching visualization? A visualization could have an interactive user interface, yet only show parts of an engine similar to an online pictorial dictionary whereas a more inquiry-based visualization might allow students to ask questions, then explore the relationships between different parts, processes, and functions; and make their own discoveries. As visualizations are designed, there may be a be a need to design multiple representations to help underscore a point and literally let people see the issue from multiple perspectives. Effectively designing interactive visualizations and ways to connect them with other learning resources to build coherence learning experience is a cyberlearning question, indeed. A second example of what we need to understand for pedagogy is: what are some effective ways to use visualizations in inquiry-based learning so learners can build their own understanding? How do we scaffold students to make inferences and understand what we intend them to learn and don’t get confused or construct incorrect knowledge?
A third set of challenges arise from the tools needed to help in the creation of visualizations. Though new cyber-tools allow us to create visualizations more quickly than they used to allow, the workflow and process of creating intelligible visualizations still needs a more integrated set of tools. Currently, there are multiple incompatible systems for accessing, merging, and transforming tools and techniques developed for professionals. Students can’t easily create their own visualizations; in fact, most teachers can’t either. More powerful tools are also needed to help with cleaning up data sets; tools are needed to do this for data mining as well as for visualizations. As we improve tools for cleaning up data, it will benefit both cyberlearning areas greatly.
Other important questions future research should address include:
In building the next generation of media literate citizens, what are some ways to teach students how to interpret visualizations, make sense of them, and attend to the principles that were used to design them?
How do visualizations complement other information sources? How do learners combine what they see with what they read, what they hear, and what they touch?
What is the power of visualizations? Under what circumstances and using what kinds of representational forms might visualizations be used to inspire citizens to change their current behaviors and possibly their belief systems? For example, visitors to a science center can interactive with a timelapse 3D data visualization of a lake system and become more aware of environmental threats to lake clarity (see Yalowitz, 2010), and subsequently take actions towards better stewardship.
Counter Point
There is indeed a lot of important research that has been done on learning from visualizations in multiple research communities including cognitive psychology, HCI, scientific visualization, and Information visualization. Yet there is still a need for more empirical evidence that modern visualizations (data visualizations and scientific visualizations) enhance student understanding and that their learning potential goes beyond their motivational value as pretty artistic images that are visually appealing.
Research in science, technology, engineering, mathematics, and medicine are moving at a rapid pace. How will fundamental research about learning and design keep up with the increasingly powerful and sophisticated tools for creating those visualizations: stereoscopy, displays ranging from 1-in to 5 meters, CAVES, AR, and other technologies?
The current way in which visualizations are produced require intense processes of gathering data, building and refining models, and adding annotations and controls. Tools like Google Earth are still too difficult for a school teacher to learn how to manipulate and use effectively for teaching without a huge investment of time. In addition, many school computers are currently too slow to take advantage of interactive visualizations.
Visualizations may be useful to convey large amounts of abstract information, but they may also obscure or oversimplify important features of models. In the worst case, visualizations become another form of lecture material that is intelligible to the expert instructor, but incomprehensible to the novice student.
Top Related Areas in Cyberlearning
Educational Data Mining
Simulations
Virtual Reality
References
Ainsworth, S. (1999). The functions of multiple representations. Computers and Education, 33, 131-152.
Ainsworth, S.E. (1999) Designing Effective Multi-representational Learning Environments, ESRC Centre for Research in Development, Instruction and Training, University of Nottingham, technical report number 58.
Chang, H.-Y. (2008). Review of Research on Dynamic Visualizations in Science Learning. In Linn, M.C., Quintana, C., Chang, H.-Y., Shen, J., Chiu, J.L., Clark, D., Menekse, M., D’Angelo, C., Schleigh, S., Touchman, S., McElhaney, K. W., Varma, K., Price, A., & Lee, H.-S. Improving the design and impact of interactive, dynamic visualizations for science learning. Proceedings of the 8th international conference on International Conference for the Learning Sciences – Volume 3. Utrecht: The Netherlands.
Gordin, D. N., Edelson, D. C., & Gomez, L. M. (1996). Scientific Visualization as an Interpretive and Expressive Medium. In D. C. Edelson & E. A. Domeshek (Eds.), Proceedings of the International Conference on the Learning Sciences, July 1996, Evanston, IL, (pp. 409-414). Charlottesville, VA: AACE.
Kali, Y., & Linn, M.C. (2008). Designing Effective Visualizations for Elementary School Science. The Elementary School Journal, 109(2), 181–198.
Larkin, J. H. & Simon, H. A. (1987). Why a diagram is (sometimes) worth ten thousand words. Cognitive Science, 11, 65–99.
Mayer, R. E. & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia learning. Educational Psychologist, 38, 43–52.
Mayer, R.E., and Moreno, R. (2002). Animation as an aid to multimedia learning. Educational Psycholology Review. 14, 87–99.
Moreno, R., & Mayer, R. E. (2007). Interactive multimodal learning environments. Educational Psychology Review, 19, 309–326.
Tversky, B., Morrison, J. B., & Betrancourt, M. (2002) Animation: can it facilitate? Int. J. Human-Computer Studies. 57, 247–262.
Yalowitz, Steven (2010). Research Study for Visualization as a Tool in Informal Science Education at Lake Tahoe (UC Davis, TERC). [Institute for Learning Innovation] UC Davis Tahoe Environmental Research Center.