Unveiling fine-scale urban third places for remote work using mobile phone big data

Third places offer a creative alternative for both work from traditional office and home, which are becoming increasingly popular. Previous studies primarily focused on qualitative analyses and survey investigations, lacking quantitative studies exploring remote work in third places. In this study, we proposed a quantitative approach to identify and characterize the fine-scale third places for remote work, with the application in Beijing, China. Initially, we identified knowledge workers who were capable of remote work through mobile office app usage. Subsequently, we delineated the finer-scale distribution of third-place visits of remote workers using mobile phone signaling data and geospatial information. Finally, we utilized the eXtreme Gradient Boosting model and SHapley Additive exPlanations value to explore the association between third-place visits for remote work and the surrounding built environment. The results revealed that (1) approximately 61.43 % of total employees had the potential to work remotely, with 11.27 % opting for remote work in third places and 4.35 % choosing specific commercial third places; and (2) the popularity of these third places was characterized by highdensity mixed-use surroundings, proximity to residential communities, and easy accessibility to subway stations. The findings can reinforce the establishment of urban design guidelines for third places, thereby contributing to the development of hybrid work models and sustainable cities.