Linking landscape features of urban parks to visitors' sense of place: A novel hybrid machine learning approach

Abstract

Sense of place is an essential aspect that enhances people’s willingness to revisit, satisfaction and loyalty. Many scholars have conducted research on people's sense of place in public open spaces. However, limited attention has been given to the comprehensive exploration of the physical environmental factors that these studies suggest affect the sense of place. Moreover, the limitations of existing traditional methods and big data approaches remain unresolved. This study aims to investigate the relationship between the physical environmental factors and tourists' sense of place, drawing on dual-function parks in Shanghai as an empirical case study. Natural language processing (NLP) and grounded theory techniques were used to establish a framework for a corresponding sense of place and its associated environmental factors in urban parks via online reviews, and word banks were created to quantify the sense of place. Regression analysis was used to determine the relationship between the sense of place, positive sentiments and environmental factors. The results showed that there was a significant relationship between the positive sentiments of park visitors and many elements of the sense of place. Among environmental factors, 12 factors were found to significantly influence the strength of sense of place. This study contributes to expanding the conceptual framework of sense of place and introducing a new approach to investigate the relationship between environmental factors and tourists' sense of place, and proposing implications for park planning and design.