Seeing is believing. Or is it? Two new researchers at the iSchool suggest there may be more going on than meets the eye in the data-driven visualizations we rely on daily – the treemaps and flowcharts, scatterplots and interactive timelines that make a mass of data accessible and digestible. A hundred questions can lurk beneath the colorful surface. Who created the visualization? What assumptions went into the design? What biases? How accurate are the sources? How reliable the presentation?

Too often, consumers don’t ask such questions when encountering an image, say Jessica Hullman and Jaime Snyder, research pioneers bringing fresh, distinct perspectives to the rapidly expanding field of information visualization studies. “There is a perception that images have a higher truth value than text,” says Snyder. “We do not question images as much as we could or should.”

And when people see something visually, whether it’s a company-generated financial chart or a newspaper graph on racial demographics, they tend to think they understand it, says Hullman. It’s easy. It’s fast.  And it may be misleading. “Visualizations can be used to conceal as well as to illuminate.”

They come at the research from different angles, with different methodologies, but both assistant professors are working to build critical awareness of the visualization process through their research and teaching. “There are people who will try to deceive with data in various ways, and it’s unlikely that I can stop that. But one thing I can do is to teach students how to recognize it when things don’t add up in a visualization, and use my own research to call attention to our biases as perceivers and as designers,” says Hullman.

Research into information visualization is much-needed in a data-glutted society increasingly reliant on what it sees for what it thinks. “The power of visualization is that you can see things you couldn’t see otherwise, whether it is patterns in large data sets or making comparisons between different types of data. It has become a very integral part of how we work with data,” says Snyder.

Hullman and Snyder join a cadre of eight new hires at the iSchool adding expertise in cutting-edge fields such as information management, curation, and visualization. “There is a growing and wonderful contingent of faculty members here interested in visualization work, expanding the course offerings in that area, and creating a path for students to explore visualization from an information point of view,” says Snyder.

Both she and Hullman come to the information sciences from artistic backgrounds. Snyder holds a Master of Fine Arts degree in visual art from Stanford University, along with a Ph.D. in information science and technology from the iSchool at Syracuse University. More recently, she completed a postdoctoral research appointment at the Interaction Design Lab at Cornell University. Much of her artistic work has focused on art installations. “Visual art is my mother tongue,” she says, “and I think I will always have an accent.” Hullman’s Master of Fine Arts degree is from the Jack Kerouac School for Disembodied Poetics, with a focus on experimental poetics and prose. She earned a Ph.D. from the University of Michigan School of Information in human-computer interaction and information visualization. She was recently a postdoctoral fellow at the University of California Berkeley in Computer Science, funded by Seattle-based Tableau Software, which specializes in information visualization products.

The new faculty members have different approaches to visualization research. Snyder takes a broad view of the field, looking at the social and ethical aspects of visualization practices in collaborative contexts – whether it’s a social media post, whiteboard presentation, or a complex data analysis. Her doctoral dissertation, which won a 2013 iSchools Doctoral Dissertation Award, looked at how the “mark-makings” on a napkin can feed higher-level communication in a social setting: A conversation’s going nowhere, someone scribbles on a napkin, and the conversation takes off, and even as the napkin is tossed, the ideas continue to fly.

“I’m as interested in the moment an artifact was created as the artifact itself. That contextual information is highly relevant to understanding meaning,” says Snyder. She points back to an early art form as an example. “I can’t imagine anyone studying cave painting who wouldn’t love to actually see the image being created.”

She has also turned her research eye on visualization in the health field. One project involved studying students who assessed their own stress levels with a feedback machine called “MoodLight.” Changing colored lights showed when they were stressed and when, through breathing or mindfulness, they relaxed – a visualization of an internal state. “I was interested in how representations of the self are provided through technology, and how that influences how we think about ourselves and each other,” says Snyder.

Next up for Snyder up is a project exploring visualization practices in citizen science projects on the UW campus. “I want to understand how visualization and representation play a role in the context of people with varying expertise and background.”

Snyder can thank her brother for pointing her into information science. He convinced the MFA graduate she should learn HTML, by herself, by hand, in order to make a living. “I discovered I loved it,” she says. Soon she was working as an HTML programmer, moving into information architecture and interaction design, project management and production, and finally into an iSchool to earn her Ph.D. 

Her colleague Hullman discovered the joy of coding while pursuing a master’s in information analysis and retrieval. She, too, was quickly hooked.

Her research work focuses on enhancing people’s ability to perceive, make sense of, and communicate with information visualizations, with a keen look at how statistical graphics tell stories around data. She studies the process in fine detail, down to preferences in sequencing: do people find the first or last thing they see most influential? “I tend to break down the design process so we become more aware of how even small decisions impact final interpretation.”

She has worked on development of several algorithmic tools that produce automatic data visualizations, including NewsViews, a visualization system that generates interactive, annotated thematic maps based on data mined from the text of a news story. She is currently working on visual tools that can support understanding of measurements, using familiar objects to explain unfamiliar measures. People may not understand what 10 kg is, but easily grasp it when told that 10kg is twice the weight of a printer, or three times the weight of a guitar. She is creating a large-scale database of object measurements that can be automatically linked to visuals for comparison.

Hullman and Snyder were drawn to the iSchool by the rich possibilities for interdisciplinary collaboration on campus and the growing number of visual information specialists in the Seattle area. Hullman plans to work with UW experts in Computer Science Engineering and Human-Centered Design and Engineering, as well as Tableau Software and Microsoft Research +.

Snyder is interested in pursuing cross-disciplinary curriculum development on campus. “I came from an iSchool with a very interdisciplinary faculty and student body. I was expected to give critiques and understand research from very different fields. So I know it takes a strong commitment to maintain that kind of work,” she says. “I wanted some place with a track record of real support for it, and I found it at the UW iSchool.”