Abstract Developing more walkable environments plays an essential role in healthy-city planning. Planners often assess the typical walkability framework from a geographic perspective based on GIS data. This paper proposes a refined walkability framework that quantifies walkability in terms of four pedestrian needs: safety, convenience, continuity, and attractiveness. Using Shenzhen as our case study, we integrate mesoscale and microscale built-environment features from different data sources: street view images (SVIs), social media, points-of-interest data, government open data, and GIS data. Deep learning semantic segmentation approaches apply to extracting street elements from SVIs. Results show that walkability and its four aspects are spatially heterogeneous in the city. Urban areas and suburban central business districts offer greater walkability than other areas. The less walkable areas are mainly in the rest of the suburbs. This work suggests that more academic and planning efforts should focus on improving walkability in less walkable neighborhoods and promoting pedestrian-friendly cities.
Using open data and deep learning to explore walkability in Shenzhen, China
2023-03-07
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Springer Verlag | 2018
|British Library Conference Proceedings | 1995
|Using Open Source Data to Measure Street Walkability and Bikeability in China: A Case of Four Cities
Transportation Research Record | 2018
|Transportation Research Record | 2003
|