Emotional polarization to conflict governance: need feature and satisfaction for aging-adapted retrofitting in historic urban communities – an IPA-BERTopic framework using social media data

Purpose China faces the dual challenges of an aging population and deteriorating old districts. This study quantifies resident satisfaction with aging-adapted retrofitting in historic urban communities to provide targeted strategies, promoting age-friendly communities and improving the quality of life and well-being of elderly residents. Design/methodology/approach User comments from five major social media platforms (e.g. Weibo and Douyin) were leveraged. BERTopic topic modeling and importance-performance analysis were applied to systematically analyze the characteristics of elderly needs and satisfaction patterns within aging-adapted retrofitting projects. Findings The aging-adapted retrofitting needs follow a “dual-peak dominance–long-tail” pattern, with aging-adapted home environments and daily services and culture at the core. Medical and health care services, along with community canteens, fall into the “advantage zone”, reflecting effective responses to core elderly needs. In contrast, indoor space planning, facilities and equipment and parking capacity fall into the “weak zone”, revealing a mismatch between expectations and actual experiences. Basic issues such as structural waterproofing and property management show low satisfaction, while community activities and smart aging services, though well-received, lack sufficient attention. Projects such as elevator installation exhibit emotional polarization, indicating deep-seated conflicts over space, safety and equity. Originality/value This study advances a replicable, scalable data-driven framework for diagnosing pre-renovation needs and supporting post-renovation monitoring. Mining unstructured social-media text reveals latent, heterogeneous and evolving demand profiles beyond preset options, enriching the differentiated-care paradigm. It converts spontaneous discussions into quantifiable topics and sentiments to guide prioritization and resource allocation, shifting aging-adapted retrofitting from experience-to evidence-driven.