Big Data-Driven Research On The Interaction Of Human Mobility Pattern And Urban Spatial Structure

Urbanization is an inevitable step in the development of modern society.The current rapid urbanization process has led to a rapid influx of people into urban space.Due to the convenient transportation facilities,mobility patterns of urban residents become complex and diverse,which is a great challenge to the stability and development of the existing urban structure,causing a series of problems of interactions between people and cities.Exploring the interaction and influence mechanism between people and cities has become a hot topic in many disciplines including urban geography,behavioral dynamics,and urban informatics.However,due to the lack of human tracking data,although the research on human mobility,activity behaviors and urban structure patterns has produced many significant achievements in their respective fields,there is still lack of knowledge on how human travels and activities rapidly change and being constrained by urban spatial structure.With the availability of vast amounts of data,interdisciplinary research has replaced traditional sparse data sources,improved traditional research paradigms,revisited existing scientific issues,and answered new questions.The growth of data and changes in demand have brought new opportunities for urban research,and a big data-driven,human-oriented urban science is emerging.Specifically,the work of this dissertation is carried out in the following four aspects:1.For the problem of individual trajectory reconstruction and activity semantics labeling,a forward-based tagging algorithm based on spatiotemporal constraints is proposed.Aiming at the issue of insufficient activity characteristics,social media check-in data is imported for feature learning.A Spatio-Temporal Constrained Hidden Markov Model(STC-HMM)is proposed.Combining the symbiosis between location changes and activity transitions,the activity semantics of the stay sequence are inferred.The algorithm experiment was carried out by using largescale mobile phone positioning data.By aggregating identified human activities,the spatio-temporal characteristics of urban activities are revealed.2.Based on the empirical massive mobile phone data,an abstraction model of trajectory network is proposed,which abstracts the space-time trajectory into a limited set of location-based motifs and activity-based motifs with locations,activities and behaviors elements.Based on statistical analysis,motif choices behaviors are further quantified.The results show that human travel seems chaotic,but it has highly regularity and preference characteristics;it proves that the“the least effort principle”surfaces in people's daily travel decisions and becomes the driving force for individuals to choose their travel structures.3.Based on large-scale human activity data,the impact mechanism of human activities on the urban functional structure is investigated.Based on the space-time clustering method,urban functional structures from activity perspective are inferred and diurnal dynamics of functions are revealed.The results indicate that the real urban functions are different from the traditional land use;the urban areas provide different activities over time,and the urban functional structure has spatio-temporal dynamics of diversity,uncertainty and diurnal transitions.4.Based on large-scale traveling data,the impact mechanism of urban spatial resources on the traveling network structure is investigated.By analyzing topological characteristics of traveling networks,the distorted characteristics of traveling networks due to the uneven distribution of urban spatial resources are revealed.The results indicate that the distribution characteristics of urban resources lead to real traveling network is neither so regular nor so random,but exhibits a hierarchical structure;the traveling networks with different complexities greatly distorts at different degree;the network with higher complexity has a larger distortion degree.