Research Topics

Remote sensing uses satellite sensors to observe the physical characteristics of the Earth’s surface; social sensing treats human beings as individual sensors and their digital traces as the spatiotemporal signals, offering a transformative perspective for understanding the socioeconomic life of cities (Liu et al., 2015). 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 social sensing 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. 

Comprehensive social sensing enables us to examine emerging and classic geographic principles through rich, dynamic, and spatially explicit evidence, such as the cosmopolitan mixing hypothesis (Nilforoshan, 2023), the 15-minute city (Moreno et al., 2021), central place theory (Christaller, 1933), and urban expansion models (Batty, 2008). In GeoDI, we advance the framework of comprehensive social sensing to understand urban geographic systems through connected places, network structures, and evolutionary processes. By comprehensively profiling the multifaceted interactions between socioeconomic activities and physical environments, we aim to provide actionable place-based insights for the digital twins of cities, data-driven planning, and policy-making toward societal good.
 

We advance spatiotemporal analytics and models for geographic knowledge discovery, especially spatial interpolation, regression, and spatiotemporal process understanding. 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. At the intersection of GIScience, spatial statistics, and artificial intelligence, GeoAI provides new ways to represent, measure, model, and explain such complexity. 

Our research explores new computational tools (e.g., deep learning, machine learning, graph-based models, generative AI, agentic AI, and geo-foundation models) as opportunities to unlock the potential of geographic principles, especially spatial dependence, distance decay, scale, regionalization, heterogeneity, spatial interaction, organization. Linking emerging AI techniques (e.g., representation learning, convolution, attention, transfer learning, meta-learning, and agentic reasoning) with key concepts in GIScience, we aim to develop interpretable, trustworthy, spatially-explicit, applicable, and theoretically informed GeoAI methods to tackle diverse urban and socio-environmental challenges.

Complex networks emerge and are observed in different complex systems such as cities, bioscience, transportation, finance, and society. More importantly, cities pulse with life. People and resources move through streets and neighborhoods, weaving spatial structures that sustain the health and vitality of urban environments. Yet the patterns emerging from human-environment interactions remain difficult to map, explain, and predict. Our research examines cities as complex spatiotemporal systems, particularly how places are connected, networks are organized, and evolutionary processes unfold.

These questions reflect a fundamental challenge: human-environment relationships are socially constructed, spatially embedded, and temporally evolving, requiring cross-disciplinary approaches that can understand complex structures from real-world observations. In GeoDI, we adopt big data, network science, and graph-based GeoAI to examinee cities as complex spatiotemporal systems, focusing on how places are connected, how spatial networks are organized, and how urban processes evolve across time and scale. Rooted in spatial statistics, network science, and GeoAI, this research agenda advances geographic computational models and network-informed analytics to study collective spatiotemporal patterns across three scales: places and their contexts, networks and their structures, and evolutionary processes and their mechanisms.