Smart city assessment is important for the construction of smart cities. Evaluation indicators for new-type smart cities (GB/T33356-2022) (EI4NSC) is the latest national standard for smart city assessment in China. However, the standard lacks defined calculable third-level indicators and a methodology for subjective-objective indicator integration. To address the problems, we defined calculable third-level indicators for EI4NSC, derived subjective indicator scores from Weibo data using pre-trained language model, calculated scores of third-level indicators by integrating Gray Relational Analysis (GRA) and Indicator Relational Analysis (IRA) methods, proposed a time-series clustering model to recognize smart city development patterns. We assessed the development of 297 prefecture-level cities from 2017 to 2021 in China, the results show that (1) Progressive improvement with "polarization and middle-tier consolidation" in smart city development level: the proportion of cities scoring 50 or less decreased from 23.91% to 21.55%; 50-60 increased from 54.88% to 56.90%; 60-70 stabilized around 16%; and 70 or more increased from 5.1% to 5.7%. (2) Smart city development in China exhibits a regional imbalance: strong in the east, transitional in the center, and weak in the west. Six development patterns are recognized: Comprehensive Development (3.4%), Well-equipped (18.9%), Ecological Construction (32.2%), Economy-driven (16.2%), Livelihood Optimization (12.1%), and Backward-looking Catch-up (17.2%). The eastern region is at the highest level and has diversified patterns, while the western region faces structural constraints and needs to explore localized paths. The results reveal the dynamic trajectory and the law of regional differentiation in smart cities, supporting differentiated development strategy formulation.
Chen et al. (2026) studied this question.