Data visualization is meant to help people come to grips with complex data, discover new and insightful findings, and communicate vital information to mass audiences. Yet, visualizations can fail to provide important context, can be manipulated by bad actors, or simply act as a meaningless distraction. If we are not careful, the rampant misuse and abuse of visualizations could undermine the entire project of using data to improve decision-making. In this talk I will discuss several areas of recent work that attempt to stave off this disaster: to characterize new ways of thinking about how visualizations can mislead, study the impact of potentially harmful design practices, and to engineering useful ways of detecting and alerting designers and analysts to potential perils in their charts.
Michael Correll is a lead research scientist at Tableau Research, where he works on understanding how to ethically, accurately, and responsibly communicate data. His areas of interest within the study of information visualization include graphical perception, uncertainty visualization, and data ethics. A particular focus of his recent work is "black hat visualization": deceptive or dangerous practices that can result in unethical or misleading uses of data. Prior to Tableau, he was a post doc at the UW Interactive Data Lab, supervised by Jeff Heer. He received his PhD. from the University of Wisconsin-Madison, advised by Michael Gleicher, where his dissertation focused on using visual perception to improve statistical graphics.