Analyzing emission and carbon reduction support policies using latent Dirichlet allocation and a Sankey-bubble chart

This study presents an in-depth analysis of China’s emission and carbon reduction support policies from 2016 to 2023 using text mining techniques. The main objective is to examine the evolution, thematic focus, and implementation outcomes of these policies across different stages, thereby providing insights into their development patterns and potential future direction. Based on the latent Dirichlet allocation model implemented in Python 3.7, the study identified and refined 14 initial topic terms spanning three policy phases, which were subsequently integrated and interpreted. Through topic clustering and visualization using the Sankey-bubble chart, the research simulated the evolution of policy themes over time. The results reveal a clear shift in policy focus – from market-driven mechanisms to green development and technology-led approaches. In the later stages, policies exhibit more comprehensive and systematic characteristics. In conclusion, the study contributes to a deeper understanding of the development trajectory, orientation, and implementation effectiveness of China’s carbon reduction policies, offering valuable insights for future policy development.