
ALT Researcher Josh Nguyen recently presented "Mapping Topics in 100,000 Real-life Moral Dilemmas" at the International Conference on Web and Social Media (ICWSM) hosted by the Georgia Institute of Technology in Atlanta. The data identifies 47 meaningful topics that fall into five meta-categories: Identities, Aspects, Processes, Events, and Things. “Aspects,” for example, includes such topics as hygiene, manners, and safety while “Events” includes weddings, breakups, and parties. The paper explains that clustering algorithms were designed to allow clusters to overlap, “since both the intersections and the gaps between two intuitive clusters (such as family and money) may be meaningful and interesting”. They found that most moral dilemmas included combinations of topics called “topic pairs”. Frequently occurring topic pairs included friends/manners, communication/mental health, and celebrations/manners. 🔗 https://cecs.anu.edu.au/100k-moral-dilemmas









