Social media data-based spatio-temporal assessment of public attitudes towards digital contact tracing applications: A case of health code application in mainland China

Digital contact tracing applications (DCTAs) are indispensable in pandemic prevention and control, but the effectiveness of DCTAs is influenced by public attitudes. Existing studies on public attitudes towards DCTAs have predominantly relied on survey-based approaches, but the survey data focused on confined time or regions has limitations in revealing the dynamics of public attitudes. This study applies a social media data (SMD)-based DCTAs-oriented public attitudes assessment approach, which includes SMD collection and preprocessing, sentiment analysis model, topic classification model, and model training and evaluation. The approach is validated with 501,095 pieces of SMD about the Health Code Application (HCA) in China, revealing the spatio-temporal evolution of public attitudes throughout the entire lifecycle of HCA. The results show that public sentiment and concerns fluctuate over time and vary across regions. There is a significant correlation between the number and proportion of different sentiments and the severity of the pandemic. Based on the level of concern, the twelve DCTA-related public concerns are divided into four categories, including high concern, medium concern, low concern, and exceptional concern. This work can offer valuable references for the formulation and enhancement of DCTAs implementation strategies in future public health crises.