Using mobile phone big data and street view images to explore the mismatch between walkability and walking behavior

Abstract

Stimulating more citizens to walk plays an essential role in building a healthy city. This paper explores the mismatch between walkability and walking behavior, using mobile phone data, street view images, and various sources of open data. Using Shenzhen as our case study, we identified walking trips of 6 months in 2021 from cellular mobile data, taking the rule-based heuristics approach. We collected ground truth GPS data to validate the walking trip extraction method. Open data and deep learning enabled quantifying walkability from the perspective of four pedestrian needs: safety, convenience, continuity, and attractiveness. We employed geospatial techniques to identify the mismatch areas between walkability and walking behavior in the city. We also explored the spatially varying effects of walkability on walking behavior. Our results showed that the mismatch areas with high-level walking trips but low-level walkability mainly occurred in the fringe areas of the central business district (CBD) and subcenters that require prioritizing more interventions. Moreover, walkability showed strong effects on walking trips in the inner suburbs. For the four aspects of our walkability framework, safety and convenience had greater positive effects on walking trips in suburbs than in urban areas. Continuity promotes walking trips mainly in the city’s western sector. The positive effect of attractiveness on walking trips clustered in the central and western parts of the city. Based on the findings, we provide prioritized and contextualized built-environment intervention strategies and policy recommendations for urban designers and transportation planners.