Sensing Urban Structure And Dynamics With Mobility Big Data

To facilitate efficient and effective city management,it is important for urban authorities to understand the regular functionalities of urban areas and the irregular crowd dynamics moving around the city.On one hand,urban planning,construction,and development have led to regular crowd movement patterns and structures to urban areas.Meanwhile,the occurrences of certain urban events may break the regular crowd movement patterns in different areas of the city.Due to the lack of understanding of the urban structures and the dynamics of urban events,urban authorities cannot evaluate the impacts of urban events effectively in the city,which hinders the short-term event management and long-term urban planning.In this paper,we leverage large-scale mobility data collected from vehicle GPS devices to analyze the urban structure in a fine-grained manner.We can sense the scope of urban dynamic changes caused by urban events based on this data.The research contents can be conclude as follows:Firstly,urban regions are divided into fine-grained areas and clustered according to similarity of crowd movement characteristics.The main steps are as follows:the urban area is meshed by the features of urban geographical area,then the taxi GPS trajectory data is mapped into the corresponding area grid and the spatial-temporal characteristics of the daily flow in a single grid is extracted.The distance-constrained clustering algorithm(DCCA)is used to cluster the regional grids according to their structural similarity and then the flow characteristics of the clustered areas are extracted.Secondly,detect abnormal traffic flow in clustered areas and interpret urban events associated with abnormal flow points.ARIMA anomaly detection algorithm is used to detect peak points of abnormal crowd based on the daily flow characteristics of urban crowd moving from the clustered area.Urban events are interpreted for abnormal flow points,and each abnormal point is associated with urban events,then collation into the database of urban event impact scope.Identify the impact of different types of urban events.We can determine the scope of urban events and use visualization techniques to demonstrate the impact of different types of urban events according to the detailed records in the urban event impact database.In this paper,we use the urban crowd mobility model to clearly describe the urban structure,and analyze the impact of different types of urban events on urban dynamics in depth.The results of these analyses are displayed in a visual display interface.Provides decision-making support for city managers to ensure the regularity of urban life.In addition,the ability to perceive urban structures and dynamics based on crowd movement patterns is enhanced,which is conducive to the rapid processing of urban events and the smooth planning of urban planning.