Dissertation:

The interaction of

research goal,

data type, &

graphical format

in multivariate visualization

All graphics and papers in these Web pages are copyrighted by Dr. Yu, Chong-ho

ABSTRACT

  • Problem: While physical and engineering scientists endorse the use of high-dimensional graphics for data analysis, several psychologists have been unable to confirm the effectiveness of three-dimensional graphs.

  • Framework: To explain this discrepancy, this study proposed an alignment framework maintaining that a successful data visualization results from the proper combination of data, task, and graph types.

  • Hypothesis:

    • Based upon the alignment framework, it was hypothesized that under conditions of medium and large data sets 3D graphs would outperform their 2D counterparts for the tasks of examining relationships and spotting outliers.

    • Also, it was expected performance would not vary across graphical formats when small data sets were used.

  • Findings: Twenty-three graduate students with experience in data visualization participated in an experiment to test these hypotheses. Superior performance for 3D graphics was found across all data sizes for both research tasks. Results are largely consistent with the theoretical expectations derived from the alignment framework.


An example movie of using 3D mesh plot to detect outliers. 



Navigation

Simplified Navigation

Table of Contents

Search Engine


Press this icon to contact Dr. Yu via various channels