근린환경 특성과 도시활력의 비선형 관계 분석 : 해석 가능성 기반 기계학습 모형의 적용

초록

Creating a vibrant neighborhood environment is a key component of sustainable urban development. Urban theorist Jane Jacobs explains that urban vitality occurs through the interactions of human activities with neighborhood environments. Drawing on the recent development of big data and machine learning technologies, this study analyzes the impact of neighborhood environmental factors on urban vitality. This study utilizes big data such as De Facto Population, Points-Of-Interest (POI), and Street View images for the city of Seoul and employs a machine learning model to understand urban vitality. It derives key variables that affect urban vitality and checks the nonlinear relationships between variables by utilizing explainable machine learning model. The main analysis results are as follows. It also indicates that land use characteristics and POI show strong associations with urban vitality. Specifically, SHapley Additive exPlanations (SHAP) analysis results confirm that the independent variables largely show nonlinear relationships with urban vitality. Moreover, the study identified critical thresholds for variables such as residential area density and distance to subway stations, beyond which their impact on urban vitality becomes constant. This study is significant because it provides a clearer understanding of the key neighborhood environmental factors that affect urban vitality. Furthermore, this study offers planning and policy implications that promoting urban vitality and social interaction.

키워드