基于移动大数据的城市空间结构感知

准确描绘城市结构、深刻理解城市特性对于智慧城市建设、提升居民生活体验至关重要。遍布于城市的基础设施,如公交车站、银行网点、移动网络等,提供了大量具有时空标记、描述个人行为的数据,使得基于人类行为的城市时空特性分析成为可能,且逐渐成为城市研究的热点。针对这些移动大数据,提出一种方法,将人类移动行为投影到城市空间实体之间的联系中,从而可以方便地从整体和局部研究城市空间实体间联系。基于该方法,利用中国西南某地级市手机上网数据进行实证研究,发现城市空间实体间联系的变化具备周期性,且城市空间结构存在与行政区划相一致的区域中心,这些区域中心的时空变化,反映了人类作息习惯与移动特性。研究结果揭示了城市空间联系的动态特性,为数据治理城市、基础设施的合理配置和资源分配提供了有效的依据。所提出的方法提供一个新的视角,将人类运动整合到空间分布式系统的研究中。

Urban Sensing based on Mobile Big Data

Depicting accurately the spatial structure of cities and deep understanding urban characteristics are essential to construction of smart city and improvement of the inhabitant quality of life. Pervasive infrastructures in cities, such as bus stations, banking outlets and mobile network, enable to capture large amounts of human behavioral data with temporal-spatial label, which makes it possible to analyze the spatial structure of cities and their dynamical properties, triggering a renewed interest for the study of urban dynamics. Based on these mobile big data, a method incorporating human movement behavior into relationships between spatially distributed infrastructure is proposed, resulting in mining ties among urban spatial entities from global and local perspective conveniently. Making use of proposed method, empirical analysis is based on data usage detail records from the cellular network in a city in western China, and the results indicate that relationships among spatially distributed entities change periodically and there are regional centers coincident with administrative division existing in the spatial structure of cities, and temporal-spatial variations of these regional centers reflect human schedule patterns and mobility characteristics. In this article, our findings reveal dynamic nature of the spatial structure inside a city, and the proposed approach provides a new perspective on integrating human movement into the study of a spatially distributed system.