我国分布式光伏工程舆情主题分布及情感特征分析——以抖音平台为例

Analysis of Public Opinion Theme Distribution and Emotional Characteristics of Distributed Photovoltaic Projects in China—Taking the Douyin Platform as an Example

摘要: 本研究爬取了抖音平台有关分布式光伏的视频数据和对应评论数据，运用BERTopic模型开展主题建模，运用StructBERT模型进行情感分析，探究分布式光伏舆情的主题分布与情感特征。研究发现，民众关切覆盖评估决策、安装、运维全生命周期的15个细分主题，评估决策阶段关注度最高；舆情整体呈现“关注度与负面情感正相关”规律，各阶段情感特征差异显著，评论情感还存在华北、华中正向领先，东南沿海、西北负面集中的空间梯度分化，且平台舆情以负面传播为主、正向科普关注度低。基于研究结果，本文从规范市场秩序、强化风险防控与区域差异化发展、健全舆情治理与正向引导三方面提出政策建议，为分布式光伏产业高质量发展提供参考。

Abstract: This study crawls the video data and corresponding comment data of distributed photovoltaic on the Douyin platform, uses the BERTopic model to carry out theme modeling, and combines the StructBERT model to conduct sentiment analysis, so as to explore the theme distribution and emotional characteristics of distributed photovoltaic public opinion. The study found that public concerns covered 15 sub-topics covering the entire life cycle of evaluation decision-making, installation, and operation and maintenance, and the evaluation decision-making stage received the highest attention. Public opinion as a whole shows the law of “positive correlation between attention and negative emotion”, and the emotional characteristics of each stage are significantly differentiated, and there is also a spatial gradient differentiation of comment sentiment in North China and Central China, and negative concentration in the southeast coast and northwest frontier, and the public opinion on the platform is dominated by negative communication and low attention to positive science popularization. Based on the research results, this paper puts forward policy suggestions from three aspects: standardizing market order, strengthening risk prevention and control and regional differentiated development, improving public opinion governance and positive guidance, and providing reference for the high-quality development of the distributed photovoltaic industry.