Residents' sentiments towards electricity price policy: Evidence from text mining in social media

Abstract According to the theory of sentiment motivation, residents' sentiments are an important factor affecting residents' electricity consumption behavior. Based on 149.95 thousand microblog posts and natural language processing methods, we analyze the time-varying characteristic, seasonal characteristic and mechanism of residents' sentiments towards electricity pricy policy. The main results are as follows. (1) Although the step tariff policy leads to the rise of electricity price, residents show positive sentiments to electricity price policy. (2) The intensity of residents’ sentiments is characterized by three stages. In the early stage, residents' sentiments towards electricity pricy policy are the most negative. In the middle stage, residents’ negative sentiments towards policy gradually decreases. In the later stage, residents show positive sentiments as a whole. (3) Residents' sentiments towards policy show obvious seasonal difference. Residents' negative sentiments show the highest intensity and obvious convergence characteristic in summer, mainly around 0, while residents' sentiments towards policy diverge in a positive direction in winter. (4) Residents' negative sentiments towards electricity policy result from smart meters, electric heating, renewable energy development and electric sector. The driving forces of residents' positive sentiments towards policy include policy cognition, public participation and policy content. Social media data provides real-time feedback on policy, which is of great significance to the formulation and improvement of policy.