| Title |
Urban Decline Diagnosis and Analysis of Urban Decline Factors Using Large Language Models |
| Authors |
김이정(Kim, Yijeong) ; 이수기(Lee, Sugie) |
| DOI |
https://doi.org/10.38195/judik.2025.10.26.5.95 |
| Keywords |
도시 쇠퇴경관; 가로경관 이미지; 주관적 인식 분석; 대규모 언어모델 Urban Decline Landscape; Streetscape Images; Subjective Perception; Analysis; Large Language Model |
| Abstract |
Urban decline critically affects quality of life and local identity. In response, the Seoul Metropolitan Government’s 2030 Urban Regeneration Strategic Plan emphasizes understanding citizens’ needs, highlighting the importance of analyzing subjective perceptions of urban decline. However, foundational data on such perceptions remain scarce. This study addresses this gap by employing streetscape images and a Large Language Model (LLM) to collect and analyze citizens’ perceptions of declining urban landscapes in Seoul. Using Kakao Street View images, a perception-based survey was conducted, followed by sentiment analysis, topic modeling, and clustering. The LLM’s responses showed a strong correlation with survey results, validating its reliability, and physical environmental factors such as deteriorated buildings and lack of greenery emerged as key drivers of decline. This study empirically demonstrates the potential of LLMs in identifying human-perceived urban decline factors, offering insights for future urban regeneration strategies. |