Inferring urban functions through human mobility patterns using taxi trajectory data

Abstract

Understanding cities requires both understanding of urban spaces and the interaction between urban spaces. Urban interaction occurs in forms of human movements. Therefore analyzing urban space through human mobility can help understand human activities in urban spaces. Types of human activities in urban space determine urban functions but not by planning interventions. For example, central business district would be actively used as the core of the urban activities as planned, but it cannot be used as intended. Monitoring urban functions therefore is critical in urban planning. Human mobility can be used to reveal these human activities. Urban mobility have possibility to represent subtle changes of urban functions. Analyzing urban functions using urban mobility, is in early stage. If we understand and apply urban mobility data properly, we can measure and analyze urban phenomena. Recent availability of urban mobility data in a large scale can be used to verify urban theories and to make new theories of human behaviors in a city wide scale. This dissertation aims to answer the following key question: "Does urban mobility have potential to reveal urban function?". The following research questions are lined to answer the key question, from understanding potentials of taxi dataset to application of the mobility data to a specific urban function i.e. commercial space: (1) What are properties of mobility dataset and potentials to infer urban function? (2) Do urban mobility dataset be able to distinguish land use that is the bedrock of urban function? (3) Can mobility dataset be used to categorize urban commercial area? To answer these questions, we firstly focus on points that taxi rides frequently occurs. We suggeseted and applied new mobility measure that calculates potential of rideshare. We showed that urban mobility could reveal urban functions. Then we examiend and verified the relationship between land use and the spatio-temporal taxi ride patterns in a city-wide scale. Finally we investigated commercial functions, one of the important functions of urban space using diverse movement measures. This dissertation adds knowledge to urban mobility and urban functions regarding urban mobility patterns. The spatio-temporal patterns of urban mobility can reveal urban functions and understanding interrelationship between urban mobility and urban space advance our understanding of human behaviors in urban spaces. We demonstrated the possibilities of the mobility dataset to measure, monitor, and characterize urban spaces.