Designing and Evaluating a Discourse Analysis Dashboard
Year: 2024
Type: conference-paper
Abstract: Extracting meaningful insights from online public discourse is crucial for managing the spread of misinformation, yet remains challenging. To effectively take action and monitor online narratives, people working towards the issue of reducing the harm of misinformation typically engage in a laborious process to make informed decisions in a timely manner. While various methods and tools to support this process have been explored, approaches to considering the entire discourse for context have not been widely studied due to their complexity. Individual texts are easier to process, whereas modelling context from an entire discourse is challenging and prone to information overload. We propose a human-AI collaborative approach through an interactive dashboard for analysing online social media at the discourse level. This paper presents the design process and preliminary evaluation of the dashboard. We detailed the co-designing sessions with experts to obtain a preliminary evaluation of this tool demonstrating its potential use and trustworthiness. We aim to conduct further user studies to develop a robust and trustworthy solution for managing and understanding online social media discourse.
Institution: University of Nottingham
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: gold
Award: G905227802
OpenAlex ID: W4402659089