Purpose This study introduces a data-driven approach to automatically construct a city-specific image evaluation index system by leveraging city-related social media data. Rather than focusing on a single, static framework, the research aims to develop a flexible and generalizable methodology capable of dynamically identifying and reflecting the distinctive attributes and evolving public perceptions of any city. Design/methodology/approach A 10 year dataset (2012–2022) of city-related Q&A posts from a Chinese community platform was collected. A preliminary index system, developed from existing literature, was refined through Latent Dirichlet Allocation topic modeling to integrate city-specific public concerns This process yielded a final hierarchical structure consisting of four primary dimensions and 12 secondary indices. To assign appropriate weights to these indices, a hybrid evaluation method combining expert judgment with data-driven analysis was employed. Sentiment analysis was then performed to gauge public attitudes toward each index, enabling a comprehensive assessment of the city’s image. Findings Wuhan’s overall image improved from 2012 to 2022, despite a mid-period dip linked to construction activities and economic challenges. The system identified economic strengths in employment and innovation, persistent concerns in housing and income, strong cultural identity, and gradually recovering environmental perceptions. Originality/value This study provides policymakers with a scalable, data-driven framework for monitoring city image, pinpointing public concerns, and prioritizing urban development strategies that align with the expectations of both residents and visitors.
Wu et al. (2026) studied this question.