Fine-grained evidence on how perceived visual quality (PVQ) varies within urban sightseeing routes remains limited. This Research Note presents a methodological proof of concept combining semantic segmentation of street-view panoramas, ratings from platform-recruited participants, and multiscale geographically weighted regression (MGWR) in a route-contextualized point-level framework. Participants rated all panoramas across five sessions, and the aggregated mean ratings showed good reliability. MGWR outperformed OLS and GWR and identified associations operating at different spatial scales. Sea visibility was broadly positively associated with PVQ, while buildings and signboards were generally negatively associated; associations for trees varied by location. The pipeline provides a replicable screening tool for identifying candidate locations for field investigation and route management.