| The increasingly development and sprawl of modern city has aroused impending demands in deeply understanding the urban spatial structure, exploring human movement regularities, and evaluating the mutual adaption between group activities and spatial structure. With the blossom of information technology, here comes the Big Data era, while at the same time generates huge and precise human location based information, such as trajectory data, cell phone data, etc. Comparing to traditional census data or travel survey dairy, these locating data take the advantage of covering a wide spatial range of human mobility, long period recording time, large sampling volume, and real-time available. The access to these big data bring new research directions to human activity intensity, activity space, and spatial-temporal dynamic inside the urban city. One the other hand, transportation networks play an important role as a bridge in connecting the human activity and urban space. It is crucial to study the distribution of human activity inside the transport network and explore the network spatial temporal patterns and regional aggregation characteristics associating to both the transport network and the urban geographical space, in order to uncover the interaction between the three.Many city have built the taxi-based floating car system out of the need of management and service. In the system, taxies in the city are equipped with Global Position System (GPS), which could collect coordinates and occupied or unoccupied service status of each taxi in a sampling interval. Then the data are transmitted back to the control center via wireless technology and stored in the transport database. Taxi is one of the most important traveling modes in public transport for urban citizens, especially in large metropolitan city. In cities served by a large fleet of taxies, taxi trajectories offer a representative portrayal of the life of the city from the perspective of city users. The human movement reflected from taxi trip data are widely distributed and have complex patterns, which are not only affected by the socioeconomic factors, but also constrained by space and time.This paper focuses on the movements and activities of particular group of population that reflected by taxi data (i.e. taxi passengers), studies their space time patterns of movements, analyze the similar activity space in different residential communities, mining the network spatial-temporal characteristics of the trajectories by taking the transportation network as a basic reference, and studies the community cluster structure in urban space by taking account of the human activity data. More specifically, the contents of this paper are:1. Introducing an outlier detection algorithm for spatial temporal taxi trajectory data. Taxi trajectories are featured from other trajectory data because they are constrained strictly in road network, have big sample data, and are heterogeneously distributed. Considering these features, we employ three factors including the ratio of travel length between GPS trajectory and shortest distance path, cost index of experiential avoid road, and travel time. The evidence based on these three factors are combined in evidence theory framework in order to detect the anomalous trajectories in the data set.2. Presenting a method to analyze the activity intensity and activity space based on the pick-up and drop-off information of taxi trajectories. First we present a method to match the Origin/Destination of taxi activity with the location that the activity taken place in. Then several indices are employed to describe the activity space of a land use polygon by taking account of the geographical distribution of Origin/Destination of taxi trips from people living in that residential community. The residential land use locations with similar activity intensity and activity space are clustered by an agglomerative hierarchical clustering algorithm. The differences between each cluster of resident land uses are analyzed in order to get common socioeconomic detail information about the resident communities, in order to supplement the lansduse classification.3. Proposing the definition of functionally critical network locations (FCNLs) in a transport network and studying the taxi trajectory patterns distributed in a transport network. A geospatial method is proposed to extract functionally critical network locations from people's moving trajectories based on the street network considering the topological relationship of network nodes, traversing patterns of trips, and trajectory density. A critical location network (CLN) is build based on these functionally critical network locations. We introduces ServeRatio as an index which is measured as the ratio of the serving trajectories and the number of whole trajectories involved, to depict the cumulative number of trip trajectories that are served by any nodes in CLN. Then we studied the evolution of the spatial extent and temporal variation patterns of FCNLs referred to ServeRatio change over the time span in the whole city and proposed two groups of quantified indices. The regularity and abnormality of human activities that are uncovered by the analysis of spatial temporal distribution of the FCNLs.4. Exploring the regional characteristics based on the network of FCNLs in an urban space. We detect the community structures from different networks that are constructed from the physical street network, the gridded network, the traffic analytic zone based network, and the critical location networks, in order to study the regional aggregation patterns based on people's movement data in an urban city. The characteristics of these socioeconomic clusters reflected from the community detections are compared with the administrative divisions. Furthermore, the differences and similarities between results of difference types of networks are analyzed to get the regional patterns of the whole city from different perspectives. The results would be useful information to support the urban planning and police making. |
PDF Full Text Request