Assessment Of Human Settlements Based On Network Science

As the level of urbanization in China increases year on year,human settlements are becoming increasingly complicated.Through theories of network science,spatial models with predictive capabilities can be established to help architects gain a rational perception of the spatial state at the design stage.This study explores the possibility of network science in constructing predictive models for human settlements.The study is divided into three parts:(1)revealing the characteristics of contemporary human settlements in China and expanding network science theory based on existing research.On the theoretical side,based on existing research using one of the indicators of network centrality,two approaches,the spatial network resilience analysis method and the network science-based spatial hyperlink analysis method,are proposed to expand the theoretical boundaries.In practice,the theoretical approach is translated into a network analysis plug-in,and the theory’s accuracy is verified in an empirical study.(2)Information technology-based methods are used to collect data on the environmental behavior.Based on this,a more refined model can be constructed.The purpose of this approach is to assess the accuracy of spatial network analysis theory as a yardstick.There are two types of data.One uses deep learning-based algorithms to obtain pedestrian flow and pedestrian trajectory.The other is to crawl data on the web to reflect the human environment’s spatial usage.The data types include geographic data and map panoramic image data.(3)Empirical research is used to provide recommendations for the design of future human settlements.It can be divided into two categories as follows.The first is a validation study,using Beijing as an example to confirm the validity of the network centrality analysis mentioned in the existing studies.The second is an exploratory study,using two typical commercial complexes as examples to summarise the validity and possible improvements of two new methods:spatial network resilience analysis and spatial hyperlink analysis methods.The key results of the paper are summarized as follows.(1)Confronts the shortcomings of traditional analysis methods.Based on network science,its role in analyzing human settlements is explained.(2)The development of Orchid,a network analysis plug-in of Rhino3D/Grasshopper,adds to the existing research by adding new methods and introduces the analysis process.(3)A new information technology-based approach to acquiring environmental behavior is used in this study.Not only does it significantly reduce the cost of data acquisition,but it also allows for a more refined picture of the state of spatial use,which in turn provides a scientifically valid yardstick for assessing the predictive results of network science.