Remco Chang
Tufts University
Chair: Kwan-Liu Ma
Topic: A Visualization Is Worth a Thousand Queries: Towards a Foundation for Reasoning in Visual Analytics
Date: 9:30 ~ 10:30, April 23
Abstract: As visual analytics continues to mature, so too must our understanding of its conceptual foundations. In this talk, I propose a framing of visualization in terms of functions, spaces, and grammar, offering a perspective aimed at formalizing how visualizations mediate between data, analytic tasks, and reasoning.
We begin with the notion of visualization as a function: a mapping from data and interaction parameters to visual representations. This functional view, illustrated through systems like NeuralCubes, positions visualizations as surrogates for query execution. More broadly, it invites us to consider query spaces: the range of potential questions a visualization can support, suggest, or help refine. At the heart of this framing is the concept of design-specific transformations, where visual encodings embed analytic computation directly into their form. A pie chart, for example, does more than display values; it pre-computes part-to-whole relationships, reducing cognitive effort and guiding interpretation. These often-invisible transformations shape the interpretive work a visualization performs and influence the kinds of questions it can help answer. Making them explicit allows us to more precisely describe a visualization’s analytic affordances.
Building on this, we introduce a hypothesis grammar, a conceptual framework that captures the structure of analytic questions. This grammar makes it possible to represent ambiguous or underspecified queries as modifiable query spaces. By aligning user tasks with data transformations and visual encodings, it helps enumerate the space of testable questions a visual analytics system can meaningfully support. Together, these ideas suggest a path toward a more formal theory of visual analytics. By conceptualizing visualizations as functions, spaces, and grammars, we can bring data, task, and visual design into a unified framework—one that offers a more rigorous and operational foundation for the visual analytics community.
Bio: Remco Chang is a Professor in the Computer Science Department at Tufts University. He received his BA in Computer Science and Economics from Johns Hopkins University, his MSc from Brown University, and his PhD from the University of North Carolina at Charlotte. Before his PhD, he worked at Boeing on real-time flight tracking and visualization software and later served as a research scientist at UNC Charlotte. His research interests include visual analytics, information visualization, human-computer interaction (HCI), and databases, with support from the NSF, DARPA, Navy, DOD, Walmart Foundation, Merck, DHS, MIT Lincoln Lab, and Draper. He is a co-founder of two startups, Hopara.io and GraphPolaris, and has received best paper, best poster, and honorable mention awards at VIS, CHI, EuroVis, and VDA. He served as Paper Chair of IEEE VIS in 2018 and 2019 and was the General Chair for VIS 2024. He is an associate editor for ACM Transactions on Interactive Intelligent Systems (TiiS) and IEEE Transactions on Visualization and Computer Graphics (TVCG) and was awarded the NSF CAREER Award in 2015. His former PhD advisees and postdocs now hold faculty positions at institutions including Smith College (x2), DePaul University, Washington University in St. Louis, University of Washington, University of San Francisco, University of Colorado Boulder, WPI, San Francisco State, Utrecht University, and Brandeis, as well as research positions at organizations such as Google, Draper, Meta, MIT Lincoln Lab (x2), the National Renewable Energy Lab, and Idaho National Lab.