Reflections on the Visualization for Social Good Tutorial
This post co-written by the organizers of the Visualization for Social Good tutorial at IEEE VIS 2019 in Vancouver, BC: Leilani Battle, Michelle Borkin, Michael Correll, Lane Harrison, Evan Peck, and Uzma Haque Syeda.
What We Did
Information visualization is often portrayed as just one more arm of the data science octopus, a highly technical skill that is employed to help scientists and specialists with esoteric data sets navigate their enormous databases or noisy collection of .csv sheets. Amongst academics, there’s sometimes an impression that persuading or advocating with data visualizations is somehow beneath us, that this sort of deviation from “just” presenting the data is the job of infographic designers and propagandists rather than “serious” researchers. But we have a great deal of power as technologists, and it’s up to us to make sure that we use this power responsibly.
While there’s a temptation to think of ourselves on the academic side of visualization as neutral, merely presenting the data in as objective a fashion as possible, often there’s nothing neutral about the data we’re presenting, no matter how much we might wish to stay on the sidelines. Who gets to collect the data, who is monitored or measured, how data persuades and drives decision-making: all of these considerations mean that there’s an inescapable social and political component to the data we visualize.
On Monday, October 21st, 2019, we packed an upstairs conference room at the Vancouver Convention center with more than 40 attendees of the IEEE VIS conference for a tutorial. What these attendees had in common was a curiosity about how they could use their skills, passions, or interest for social good. The format of the tutorial was a central critique and redesign, interrupted by periodic lightning talks by the organizers. Our attendees self-organized into groups based on topics they were passionate about (like the media, accessibility, or data literacy), looked for visualizations around those topics, and discussed how these visualizations might be redesigned to have more impact.
We as speakers tried to give the participants a wide variety of topics to think about as they considered this exercise. Leilani Battle began with a discussion of the terminology that comes along for the ride with a concept as nebulous as “social good.” Michael Correll discussed the long history of persuasive visualizations and techniques to persuade with data. Evan Peck discussed barriers to the circulation and understandability of visualizations. Lane Harrison discussed the unique challenges in designing for social impact. And Umza Haque Syeda finished up by telling us how integrating service and learning can produce positive impacts for both students and organizations. We’ve stored the talk slides here if you’re curious.
But we had a (not-so) secret goal here, beyond talks and collaborative activities. Often, the sorts of projects we tackle in academic visualization are based on who has the grant money and the time to work with us (especially for the months or years often required to do the sort of intense cross-disciplinary work that our collaborations often require). That’s not always a winning strategy for choosing projects that we think will yield the most good! And for a conference that typically values novelty or flash, there’s not always a place for recognizing the hard, important, but not necessarily novel work of making visualizations for broader audiences, and for broader goals. So we wanted to build cohorts and energy around vital topics that we felt were somewhat underrepresented at the conference.
Our attendees self-organized into groups on using visualizations for media, for urban development, tackling inequality, raising political engagement, promoting accessibility, and building data literacy. While some of these topics showed up in the main conference, or in workshops associated with the conference, there is often a lack of consideration of using visualization for advocacy in the main conference. This might be because our domains of choice (or the domains where funding is easiest to acquire) are so narrow and specialized that a path to engaging broader audiences seems untenable, or maybe we don’t feel like we have the scientific tools to do rigorous work with visualization in this space. It’s even possible that there’s some element of avoidance here, that once we open the door to having to consider the wider social implications of our work we’ll be forced to have a lot of long-delayed conversations we’ve been putting off.
Where To Go From Here
The core of the workshop focused on specific redesign activity: what’s a visualization in an area you are passionate about, and how might you redesign it to have a larger positive impact? But the conversation often shifted to more systemic matters, rather than the specifics of what visualizations worked or didn’t work. Yes, we were interested in the design of visualizations for tackling inequality or disinformation or literacy, but we were also interested in the forces, incentives, and barriers that encouraged or discouraged activity in those areas (although we admit we’re still looking for answers). To be blunt, why does a field of energetic, socially-attentive designers not have a larger footprint in pressing social issues?
There are a lot of potential barriers here. For instance, resources: non-profits and other NGOs typically having fewer opportunities for funding academic work than governments and profit-driven companies. Or availability: it’s much easier to get our hands on data about flowers or Titanic passengers than data with immediate social impact; some of the most crucial datasets may not even currently exist. There’s also a cultural component here: many researchers feel uncomfortable or unsafe “taking a stand” with their work, or otherwise acting as anything other than entirely neutral communicators (despite the fact that often our “raw” or “neutral” data is anything but). There are real dangers in taking on these projects as well, both for ourselves and the people we set out to help: there is beginning to be a blowback against “parachute researchers” who drop in with good intentions but don’t make the kinds of collaborative or long term efforts that will result in lasting impact. Overcoming these barriers and reducing these potential harms will require time, education, advocacy, and engagement with people outside our usual sets of collaborators.
We are grateful for the time, attention, and hard work the attendees of Vis4Good put into the tutorial. Speaking with attendees during and after the tutorial was a highlight, and we learned that there may be opportunities to further elevate the perspectives and work of people using visualization for good in the future.
This leads to our questions for you:
How can we convert energy and attention into action? How can we convert interest into a stronger representation at our premier conferences? Expand our tutorial? Create a workshop? A hackathon? A secret society? What should we do next?
