ISSN : 1225-598X
This study explores discovering latent library service demands through unstructured data analysis. We analyzed 8,186 civil complaints and 1,993 YouTube comments from 2024 using sentence-BERT embedding and clustering techniques. Five clusters from complaints and four from comments were integrated into five user segments: emotional empathy, information seeking, information needs and communication, institutional utilization, and civic participation types. These segments revealed complex motivations including identity formation, sense of belonging, and civic engagement beyond simple information needs. This exploratory study demonstrates the potential for identifying latent public library demand through online data analysis, suggesting information needs-based segmentation can enable targeted service strategies beyond traditional demographic approaches.