City networks and clusters as expressed in Chinese and Japanese languages: A multiscale network analysis with language-sensitive webpage big data

Abstract

There is considerable research on world city networks (WCNs) and on their cluster properties, yet, few have compared these properties as they appear in different languages and cultures. This study introduces language-sensitive webpage big data in English, Chinese and Japanese, and develops a novel multiscale analytical framework including the weighted stochastic block model (WSBM) to detect and compare the global and grouping features in WCNs by language. The results suggest the properties of city networks and clusters are differently viewed in different languages or cultures, with a more globalized multi-cores-periphery structure in the English WCN, a more localized spiky structure in the Japanese WCN, and a double-cores-periphery structure in the Chinese WCN. This research demonstrates the importance of recognizing the differences in WCN structures as shown in different languages and the significance of mesoscale network analysis in city network comparisons.