A study on smart city research activity using bibliometric and natural language processing methods

Abstract

Smart cities have become a new urban development paradigm and draw much interest from the research community and society. Based on academic publications of smart city-related research, this study employs bibliometrics, natural language machine learning methods to analyze 10,000 papers indexed by Web of Science from 2009 to 2020. Bibliometrics results show that: (1) A total of 114 countries or regions worldwide have participated in smart city research, and China is the country with the highest amount of participation in the field of smart cities. (2) Smart city research has gone through three stages: the initial stage (2009-2012), the in-depth advancement stage (2013-2016), and the leap-up stage (2017-2020). Researchers paid more attention to urban attractiveness indicators such as sustainability in the early stage. In the later period, most of the research topics were clustered on improving the overall function of the city. Latent Dirichlet Allocation (LDA) topic model results revealed that research topics could be categorized into five aspects: policy research on the status quo of smart cities, data analysis and application, infrastructure construction, urban governance, and network security. Current research on smart city technologies mainly focuses on theoretical systems, technologies, and application fields. There is a lack of in-depth research and exploration in long-term construction and operation mechanisms. This research provides insight into the research status of smart city technologies and helps researchers decide on future study direction.