Qiu and colleagues turn collective memory into a spatially operable urban variable without reducing it to heritage inventory. Their central proposition is that geotagged social-media traces can reveal how historical continuity, public identity and the built environment are co-produced across the city. Methodologically, the study combines collective-memory theory with Kevin Lynch’s urban-image framework, natural-language processing and GIS to construct memory maps organised through paths, nodes, landmarks, districts and edges. Fine-tuned language models classify large volumes of public text, while spatial aggregation reveals concentrations, differentiation and relations among historical, natural, educational and infrastructural memory elements. The result is a quantified urban memory field rather than a catalogue of monuments. The paper bridges heritage studies, urban morphology, computational social science and planning by treating public memory as both semantic and spatial data. Its methodological importance lies in making representation mappable while preserving the idea that memory is relational, layered and anchored in everyday urban perception.
