Research Topics

The rise of spatiotemporal big data, e.g., mobile phone data, street view images, remote sensing imageries, social media data, and smart card data, has revolutionized our ability to study urban dynamics at unprecedented scales and resolutions. These datasets enable the study of individual and collective human digital footprint patterns through multi-sourced observations such as origin-destination (O-D) flows, social media texts, vision perceptions, socioeconomic environment, urban land use, infrastructures, and longitudinal tracking of events in cities. Historically, our understanding of cities was largely theoretical or constrained by sparse empirical data; today, new sciences based on the flow networks of human mobility—formalizing collective movement as spatially embedded network structures—allow us to revisit classic urban theories by comprehensively sensing the multi-facet of cities, such as the cosmopolitan mixing hypothesis (Nilforoshan, 2023), the 15-minute city (Moreno et al., 2021), central place theory (Christaller, 1933), and models of urban expansion (Batty, 2008). Comprehensive sensing of our socioeconomic environment and how it interacts with the physical surroundings, can not only refine our understanding of urban systems but also provide foundational insights for the development of digital twins of cities, facilitating data-driven urban planning and policy-making toward societal good.

Human & Social environment

Human & Physical environment

Geographic phenomena are considered complex due to the heterogeneous nature of spatiotemporal dependencies. It is impossible to specify a universal law described in statistical or physical languages that can perfectly characterize a real-world geographic process and explain how it forms certain observed patterns. As an emerging field at the intersection of GIScience, statistics, and AI, GeoAI offers a novel and bold perspective on revisiting and advancing current spatiotemporal methods and models, with the potential of new representation, measurement, analytics, and models to cope with diverse urban challenges. Despite many efforts that utilize deep learning (e.g. CNN, GCN, LSTM, GAN, Transformer, LLM) for spatiotemporal computational research, not many exist to investigate and discuss the intuitive link between key concepts in spatiotemporal analysis and modeling (e.g. dependence, organization, scale, regularity, distance, heterogeneity, step-changes, etc) and AI principles (e.g. convolution, representation learning, attention, transfer learning, meta-learning). 

Complex networks emerge and are observed in different complex systems such as cities, bioscience, transportation, finance, and society. In the urban context, people navigate cities by responding to perceived opportunities and environmental changes, selecting destinations based on a complex interplay of spatial, social, and economic factors. Cities, as dynamic human-environment systems, evolve rhythmically, where urban infrastructures function as structural backbones, while human movements act as the circulatory system, sustaining and transforming the urban fabric. Fascinating phenomena, such as spatiotemporal scaling and fractal geometry, emerge as indicators of urban complexity. In GeoDI, we adopt big data, network science, and graph-based GeoAI to approach the complexity of cities.