Research on the spatial characteristics of sky gardens based on networked pictures: a case study of Singapore

Abstract

ABSTRACT As alternative public spaces in vertical public realms, sky gardens have become an increasingly important spatial element and part of the urban vocabulary in high-density environments worldwide. However, the public's spatial experience, social behavior and aesthetic needs in sky gardens have not been thoroughly explained. This paper takes Singapore, which has a rich social public space in the air, as an example where studies estimating spatial perception have used networked images instead of user preference survey results. Manual image identification and image recognition were performed with a computer vision application programming interface (API) to recognize networked images. Then, we extracted the sky gardens’ main elements in Singapore’s urban environment using clustering analysis, principal component analysis and factor analysis with SPSS. Finally, we discuss the spatial distribution, role and composition of sky gardens in high-density cities and people’s emotional preferences. The study shows that sky gardens are effective alternative public spaces in Singapore's vertical city, and their physical elements occupy a dominant position in the public’s spatial perception; there is a strong tendency for different buildings’ sky gardens to have spatial demand characteristics, while public facilities and cultural activities are key factors that affect sky gardens’ quality.