Research Interests
As data continues to grow in size and complexity, how will interactive data analysis continue to create insights? My primary research interests address this challenge through combining techniques from visualization, human-computer interaction, statistics, machine learning, and data mining. My students and I invent novel user interaction techniques for people to explore data through visual analytic systems. See more at our VA.lab site.
Select Projects
 |
Detecting and Mitigating Biased Analysis Strategies People are critical to human-in-the-loop visual analytics. Systems learn important domain expertise from people. However, how can we build systems that also alert people of potentially biased strategies? |
 |
Collaborative Visual Analytics Showing collaborators what parts of the data has been explored by who helps people cover more data, maintain situational awareness, and recall the process. |
 |
Visualization by Demonstration User interaction for visualization that lets people demonstrate intended visual mappings, data transformations, and visual representations to explore data. |
 |
Semantic Interaction for Model Steering Coupling exploratory user interactions with analytic model parameterization increases the usability of visual analytics. We learn about users from their interactions, and adapt the computation accordingly. |
 |
Steering Multi-Attribute Scatterplot Axes Scatterplots that show outputs of dimension reduction models suffer from interpretability and interaction complexity. Interaxis is a visual analytic technique that makes that easier. |
 |
Novel Display and Interaction Modalities for Visual Analytics More expressive forms of user interaction enhance analytic discourse between data analysis tools and analysts. We explore how different interaction and display modalities enhance insight into data. |
 |
Analytic Provenance Data analysis consists of many steps, interactions, questions, and other activities to achieve insight. We explore methods for aiding data analysts capture, retrieve, and leverage their analytic provenance during and after analysis. |
 |
Building, Steering, and Comparing Multiple Machine Learning Models for Visual Analytics As the number of machine learning models continues to increase, people often have the need to consider multiple different models for a given task. We explore ways for users to build, steer, and compare multiple model types for a given task. |
| Visit our Visual Analytics Lab website for more! |