Spatiotemporal panel dataset of China social credit system implementation extracted via large language models

China’s Social Credit System (SCS) is a key digital governance initiative, yet empirical research is constrained by a lack of high-frequency execution data. We present a spatiotemporal panel dataset of local government credit governance in China from 2000 to 2025. The dataset is derived from over 1 million news articles collected from official government portals across 31 provinces, 306 prefectures, and 469 counties. We employed a Dual-Track Information Extraction Strategy using Large Language Models: a Silver Track (Qwen-7B) for macro-classification of governance domains and sentiments, and a Gold Track (Qwen-32B) for micro-extraction of regulatory mechanisms and lifecycle phases. These news-derived indicators are aligned with socio-economic statistics (2000–2024) via standardized administrative division codes. This granular, analysis-ready resource enables researchers to evaluate policy effectiveness, analyze government behavior patterns, and explore regional heterogeneity in China’s modernization of governance.