Affective boundaries: a computational analysis of topic dynamics and sentiment in online reactions to Seoul's Filipino domestic worker pilot programme

In 2024, the Seoul Metropolitan Government launched the Foreign Housekeeping Managers Pilot Project, employing Filipino domestic workers to address care labour shortages amid low fertility and aging. As Korea's first formal institutionalisation of migrant care work, it offers a rare chance to observe how public commenters on Naver, the country's dominant news portal, interpret and contest this change. Using a mixed-methods computational approach combining manual topic classification, LDA topic modelling, and deep learning-based sentiment analysis, we analyse 543 news articles and 20,685 comments from September 2024 to April 2025. Seven themes appeared: working conditions and wages, policy rollout, public reactions, international comparisons, migrant rights concerns, the programme as a response to care shortages and fertility decline, and alternative models. Sentiment skewed heavily negative, dominated by anger and disgust tied to perceived unfairness and contested belonging. Fluctuations tracked policy events, showing how administrative developments became political flashpoints. We use the concept of affective boundary-making, advanced as a synthesis of symbolic boundary work and affective publics rather than as a new theoretical construct, to describe how anger and disgust function as the medium through which deservingness boundaries are drawn during the pre-institutional phase of a care-migration policy.