Hi I'm Shenyu Xu
Researcher in Information Visualization (InfoVis) and Human-Computer Interaction (HCI), currently with Visual Analytics Lab at Georgia Institute of Technology, where I work with Prof. Alex Endert.
Primarily focused on the design, development, and evaluation of novel visual analytic systems.
Seeking summer research interns with a focus on Visual Analytics and HCI.
Ph.D. Computer Science
Georgia Institute of Technology
Atlanta, GA
2019 - Present
B.S. Computer Science
University of California, Davis
Davis, CA
Awarded in 2017
Working on the development and evaluation of QUESTO, an interactive visual interface that allows non-experts to formulate their preferences as objective functions to build customized effective classifiers. Conducting a between-subjects user study comparing QUESTO with a pure auto-ML optimizer - Hyperopt, to evaluate the utility of QUESTO. Submission planned for EuroVis 2020.
Designed, developed and evaluated ChartComic: a novel visual analytic tool for auto generating visual narratives in the form of data comics from a collection of charts. Operationalized established design principles into an advanced backend pipeline which characterizes charts by their properties and similarity, as well as recommends the data comic’s layout, partitioning, captioning (via textual descriptions), and overall narrative ordering. Designed and completed a within-subjects user study comparing ChartComic with a baseline system that offers manual data comic creation. Evaluated the study results with statistical null hypothesis tests. Submitted to CHI 2020.
Implemented and evaluated CoVA: a Git-based visual interactive analytic system for supporting collaborative sensemaking and reasoning. Adapted version control workflows (Git) to merge multi-user insights into a comprehensive insight graph during the process of team-based data exploration. Designed and performed a between-subjects user study comparing CoVA with a baseline analysis interface. Evaluated the study results with statistical null hypothesis tests. Submission planned for EuroVis 2020.
Conducted the design, implementation, and evaluation of Chart Constellations: a visualization-based chart summarization interface for review and investigation of team-based, collaborative analysis workflows. Extended the technique of declarative visualization and visual encoding similarity to facilitate multi-user collaboration in a visual analysis system. Designed and conducted a between-subjects user study using a think-aloud protocol and evaluated the results with statistical null hypothesis tests. Published in Computer Graphics Forum (presented at EuroVis 2018).
Link to the papershenyuxu@gatech.edu
85 5th St., NW Technology Square Research Building Atlanta, GA 30332-0760 Near Office: TSRB 335
+1 (530) 400-0128