

Student:ÌýMohan Sunkara
Date and Time:ÌýJuly 23, 2025 at 12:30 PM
Location:ÌýECSB 1st floor auditorium
Ìý
Committee Chair:ÌýDr. Vikas Ashok
Committee Members:
Dr. Michele C. Weigle
Dr. Hae-Na Lee (Michigan State University)
Dr. Michael L. NelsonÌý
Ìý
Title:ÌýENHANCING NON-VISUAL INTERACTION WITH ONLINE USER-GENERATED CONTENT
Ìý
Abstract:ÌýThe Web has become the dominant medium for our everyday activities, including communication, business, e-commerce, news, and entertainment. Consequently, the online world is experiencing an explosion of user-generated content (UGC), particularly on social media platforms and online review systems. To enable convenient interaction with huge amounts of UGC, web platforms have adopted various presentation or visualization strategies that enable users to efficiently peruse and contribute to the UGC. However, these user interfaces are primarily designed for sighted individuals, and as such they do little to assist blind users who rely predominantly on audio-based screen reader assistive technology to interact with online digital content. The extant research efforts to improve web interaction for blind screen reader users have mostly focused on improving either the accessibility of visual content such as images and videos or the usability of overall webpage content navigation. While these general-purpose solutions certainly improve the usability of UGC for blind users to a certain extent, they presently lack the ability to address the blind users' unique domain-specific usability needs associated with the different kinds of online UGC. I strive to fill this gap in this thesis, by examining and enhancing the non-visual usability of two popular kinds of online UGC - online discussion forums and online review systems. For online discussion forums, I designed and developed PView, a browser extension that enables blind users to conveniently customize discussion threads so that they can efficiently and selectively navigate posts based on their interest and preferences. In a user study with blind participants, I observed that PView significant improved the usability of comments and social media threads over the status quo visual user interfaces. Likewise, for online review systems, via a user-centered design approach, I built QuickCue -- a browser extension that helps blind users efficiently explore the information in online customer reviews, through a set of thematically-focused and sentiment-aware summaries auto-generated from the raw reviews. By extracting and presenting key information upfront, QuickCue reduces the need to sift through long, unstructured texts, thereby minimizing listening fatigue and supporting more informed decision-making. To further improve user experience in online review systems, I also devised and implemented a non-visual ‘speed-listening’ technique, namely ReviewSkim, to augment both PView and QuickCue, enabling blind screen reader users to quickly skim through UGC, akin to how sighted users can quickly grasp the essence of content through visual speed-reading techniques. User studies with blind participants showed that both QuickCue and ReviewSkim significantly enhanced usability and reduced interaction workload compared to the status quo user interfaces. In sum, this thesis lays the groundwork for designing novel assistive technologies for enabling blind users to usably interact with UGC that is increasingly permeating our everyday Web activities.