Big Models: From Beijing to the whole China

TL;DR

This paper introduces the concept of big models as a new research paradigm for regional and urban studies, leveraging big data and advanced modeling to enable fine-scale simulations across large geographical areas in China.

Contribution

It defines the concept, characteristics, and potential of big models, illustrating their application through case studies in Chinese urban and regional analysis.

Findings

Mapping urban areas for all Chinese cities

Parcel-level urban simulation demonstrated

Potential for nationwide application of big models

Abstract

This paper propose the concept of big model as a novel research paradigm for regional and urban studies. Big models are fine-scale regional/urban simulation models for a large geographical area, and they overcome the trade-off between simulated scale and spatial unit by tackling both of them at the same time enabled by emerging big/open data, increasing computation power and matured regional/urban modeling methods. The concept, characteristics, and potential applications of big models have been elaborated. We addresse several case studies to illustrate the progress of research and utilization on big models, including mapping urban areas for all Chinese cities, performing parcel-level urban simulation, and several ongoing research projects. Most of these applications can be adopted across the country, and all of them are focusing on a fine-scale level, such as a parcel, a block, or a township (sub-district), which is not the same with the existing studies using conventional models that are only suitable for a certain single or two cities or regions, or for a larger area but have to significantly sacrifice the data resolution. It is expected that big models will mark a promising new era for the urban and regional study in the age of big data.