*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title: Visualization of Textual Content from Social Media and Online Communities
Mengdie Hu
Ph.D. Candidate
School of Interactive Computing
College of Computing
Georgia Institute of Technology
Date: Wednesday, April 5rd, 2017
Time: 3 PM to 5 PM ET
Location: GVU Cafe (TSRB 2nd floor)
Committee
--------------
Dr. John Stasko, School of Interactive Computing (Advisor)
Dr. Rahul Basole, School of Interactive Computing
Dr. Alex Endert, School of Interactive Computing
Duen Horng (Polo) Chau, School of Computational Science and Engineering
Michelle X. Zhou, Juji, Inc.
Abstract
--------------
The increasing popularity of social media and online communities results in unprecedented abundance of text documents that express authentic opinion and emotional state of the masses. While these documents seem like a goldmine for researchers and analysts, gaining insights from them is not easy. Given their scale it is obviously hard to read through them. Pure algorithmic solutions also face serious challenges due to the natural complexity of human language and the often open-ended nature of analytic tasks. In this thesis, I explore interactive data visualization systems that combines the power of both human and algorithms to address analytic tasks for these documents.
I focus on two domains that sit at both ends of a wide spectrum of social media text: consumer reviews and social media posts. For both domains, I outline important characters of the text that set them apart from traditional text document. I also explore major analytic tasks related to each type of document. I discuss why natural language processing techniques fail to address some of these tasks, and propose solutions that engage the cognitive advantage of the human user. I design novel data visualization metaphors and interactions, and implement working prototypes that combine my visualization and interaction design with natural language processing techniques. User studies with both novice users and domain experts demonstrate the values of my designs.
