Date of Award

12-2017

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Engineering and Sciences

First Advisor

David W. Mutschler

Second Advisor

Jennifer Long

Third Advisor

Vernon Gordon

Fourth Advisor

Philip Bernhard

Abstract

This year, every second, five gigabytes of new data will be streamed to storage and yet less than half a percent of all this data will ever be analyzed. Much of this muted data is high-variety object data, unnoticed in a void of sorely needed tools to make it readily understandable. A software tool for visualizing streaming object data in a context-free manner can be built, and this tool would aid in finding predictor data for known events which are functions of the data. The design of the tool to support the hypothesis is presented and field results from over 500 test trials are analyzed. Score distributions from participants using the tool are compared to a random score distribution. Users did correctly identify more known event predictor data than they would have if selections were made at random. Therefore, the software tool is useful.

Share

COinS