Urban Mobility Pattern Mining Based On Data Analysis

Smart city is making full use of the information technology to process the data from all walks of life in the city and then service for them, and urban mobility pattern is the basic subject of smart city. In this thesis, the author analyzes the urban mobility patterns of Nanjing based on taxi GPS data which recorded by over 7000 taxi in two months in Nanjing. The thesis focus on map-matching (Map Matching, MM), urban traffic patterns and urban human mobility patterns. Urban traffic pattern mainly concentrates on the traffic stream and the recurrent congestion segments spatial-temporal distribution, and the trip distance distribution, the trip hotspots and the flow of human is the main content of urban human mobility pattern. The local conclusion in this thesis is the guidance knowledge for urban transportation planning, region planning and public health construction.Firstly, the thesis review the background and current situation of urban mobility patterns. And then the Open Street Map (OSM), the noise of GPS data are introduced and analysis. After that he use the local ST-Matching algorithm to match the clean GPS data in the electronic map. Secondly, the author proposes an congestion segment score model which based on the conclusion that the vehicle speed on the road with same level can be fitted by normal distribution, and completes the extraction of recurrent segment in Nanjing by introducing the congestion indicators. The result shows that Zhongshan South Road, inner circle, Erqiao South Road and Bridge highway are the bottle neck of Nanjing road network.Then, the thesis analyzes the taxi travel distance and the number of taxi trips in Nanjing at different time slots with statistical method, and find that the travel distance can be fitted with power law. The research on Nanjing hot spots which studied by hierarchical clustering based on modularity presents that the crowd focused on Zhonghuamen, Andemen, Maigaoqiao railway station at the peak period, the 1912 block in the midnight instead. And the hospitals like the People’s hospital of Jiangsu Province and Nanjing railway station act as hot spots at most of the time in the day. Finally, the thesis analyzes the human flow with grid statistical method. The result shows that the main crowd flow is from the residential area to the station at the peak time, and the evening peak is exactly the opposite, while the semantic shows diversity, at other times.