Abstract
With the advancement of the Internet of Things (IoT) and communication platform, large-scale sensor deployment can be easily implemented in an urban city to collect various information. To date, there are only a handful of research studies about understanding the usage of urban public spaces. Leveraging IoT, various sensors have been deployed in an urban residential area to monitor and study public space utilization patterns. In this article, we propose a data processing system to generate space-centric insights about the utilization of an urban residential region of multiple Points of Interests (PoIs) that consists of 190 000 m 2 real estate. We identify the activeness of each PoI based on the spectral clustering, and then study their corresponding static features, which are composed of transportation, commercial facilities, population density, along with other characteristics. Through the heuristic features inferring, the residential density and commercial facilities are the most significant factors affecting public place utilization.