Visualization and Interactive Data Analysis

Refreshments served   ABSTRACT The increasing scale and accessibility of digital data provides an unprecedented resource for informing research, business and public policy. Yet acquiring and storing this data is, by itself, of little value. Turning data into knowledge is a fundamental challenge for both computer systems and user interface research: it requires integrating data management systems and analysis algorithms with human judgments of the meaning and significance of observed patterns. In this talk, I will discuss our research attempting to address this challenge through novel interactive systems for data visualization and manipulation.  First, visual representations are regularly used to aid perception of patterns, trends and outliers in data. To aid this process, we are investigating the design of declarative, domain-specific languages for custom visualization. Our resulting languages (Protovis and D3) simplify specification and enable performance optimization while preserving an expressive design space. These systems are now widely used throughout academia and industry. Second, data analysts often expend an inordinate amount of effort manipulating data and assessing data quality issues. With our Wrangler system, users can construct data transformation scripts in a direct manipulation interface. Wrangler uses programming-by-demonstration methods to automatically suggest applicable transforms and preview their results. The end result is not simply transformed data, but a reusable program that can be run on other platforms (e.g., MapReduce) to process data at scale. Collectively, these systems contribute new approaches for improving the efficiency and scale at which expert analysts work, while lowering the threshold for non-experts.   BIO Jeffrey Heer is an Assistant Professor of Computer Science at Stanford University, where he works on human-computer interaction, visualization and social computing. His research investigates the perceptual, cognitive and social factors involved in making sense of large data collections, resulting in new interactive systems for visual analysis and communication. The visualization tools developed by his lab (Prefuse, Flare, Protovis & D3) are used by researchers, corporations and thousands of data enthusiasts around the world. His group has received Best Paper and/or Honorable Mention awards at the premier venues in Human-Computer Interaction and Information Visualization (ACM CHI, ACM UIST, IEEE InfoVis). In 2009 Jeff was included in MIT Technology Review's TR35; in 2012 he was named a Sloan Foundation Research Fellow. He holds BS, MS and PhD degrees in Computer Science from the University of California, Berkeley.