TAPRAV: An interactive analysis tool for exploring workload aligned to models of task execution

Brian P. Bailey, Chris W. Busbey, Shamsi T. Iqbal

Research output: Contribution to journalArticlepeer-review

Abstract

Pupillary response is a valid indicator of mental workload and is being increasingly leveraged to identify lower cost moments for interruption, evaluate complex interfaces, and develop further understanding of psychological processes. Existing tools are not sufficient for analyzing this type of data, as it typically needs to be analyzed in relation to the corresponding task's execution. To address this emerging need, we have developed a new interactive analysis tool, TAPRAV. The primary components of the tool include; (i) a visualization of pupillary response aligned to the corresponding model of task execution, useful for exploring relationships between these two data sources; (ii) an interactive overview + detail metaphor, enabling rapid inspection of details while maintaining global context; (iii) synchronized playback of the video of the user's screen interaction, providing awareness of the state of the task; and (iv) interaction supporting discovery driven analysis. Results from a user study showed that users are able to efficiently interact with the tool to analyze relationships between pupillary response and task execution. The primary contribution of our tool is that it demonstrates an effective visualization and interaction design for rapidly exploring pupillary response in relation to models of task execution, thereby reducing the analysis effort.

Original languageEnglish (US)
Pages (from-to)314-329
Number of pages16
JournalInteracting with Computers
Volume19
Issue number3
DOIs
StatePublished - May 2007

Keywords

  • Mental workload
  • Pupil size
  • Task models
  • Visualization

ASJC Scopus subject areas

  • Software
  • Human-Computer Interaction
  • Library and Information Sciences

Fingerprint

Dive into the research topics of 'TAPRAV: An interactive analysis tool for exploring workload aligned to models of task execution'. Together they form a unique fingerprint.

Cite this