Research on Evidence-Based Policy Making with Social Media for Haze Governance

• 1、College of Management and Economics, Tianjin University, Tianjin 300072

Abstract：The large amount of user-generated content that the public discussed and shared on social media contains rich information about public appeals, opinions, and sentiments, and it can be used as evidence for public policy making. This paper analyzed the discussion contents of haze topic in Zhihu Question and Answer Community with text mining methods, including: (1) User sentiment analysis, to make clear the user's attitudes about haze topic, in order to evaluate the urgency of the problem; (2) Discussion theme analysis, to clarify the user's appeals and opinions about haze issue, and formulate policies based on it. The results showed that Zhihu users had shown a significant negative sentiment under the haze topic, and the major concerns of them includes: the causes of haze, governance methods, and protective equipment. Based on the results, the following three policies were recommended: promoting the research of clean production technologies, providing subsidies for protective equipment against haze damage, and establishing an online communication channel between government departments and the public.

Keywords： Information Management and Information System Social media Text mining Haze governance Evidence-based policy