Mershack Okoe

Florida International University School of Computing and Information Sciences


Lecture Information:
  • June 24, 2016
  • 2:00 PM
  • ECS: 349

Speaker Bio

Mershack Okoe is a PhD candidate in the School of Computing and Information Sciences at Florida International University, working under the supervision of Dr. Radu Jianu. His research interests are in Data Visualization and Data Analytics. He received his Bachelors degree in Computer Science from Kwame Nkrumah University of Science and Technology (Ghana 2007), his Masters degree in Advanced Computing from University of Bristol (UK 2009), and his Masters degree in Computer Science from Florida International University in 2014.

Abstract

We propose to reduce the overhead involved in designing and running controlled user studies of web-based data visualizations so that evaluations can be done more easily, cheaply, and frequently to rapidly test hypotheses about visual designs. To this end, we will explore the design of VisUnit an online service that will automate the process of designing, running, and analyzing results of controlled user studies of data visualizations by leveraging crowdsourcing and a repository of benchmark data and tasks.

Controlled user studies play a vital role in data visualization research because they allow us to measure the strengths and weaknesses of different visualization techniques quantitatively, and because they provide insight into what makes one technique more effective than another. However, conducting user studies requires expertise and is time consuming. VisUnit will simplify the evaluation process by allowing visualization designers to easily configure user studies for their web-based data visualizations, deploy them online, use crowdsourcing to attract participants, collect user responses, and analyze incoming results automatically using appropriate statistical tools and graphs.

We provide preliminary evidence that this approach can be effective by showing that GraphUnit, an evaluation engine for network visualizations we recently developed, allowed us to replicate previously published user study designs in a matter of minutes.

We hypothesize that lowering the overhead involved in conducting user studies would allow user studies to be used not just to evaluate finished designs, but also to help visualization designers make informed design choices at intermediate stages of the design process. We will test this hypothesis by using VisUnit to drive the design and development of a novel web-based neuroscience visualization system which supports the workflow of a neuroscientist.