Exploration of Social Media Opinions on Innovation for Sustainable Development Goals by Topic Modeling and Sentiment Analysis

This research utilizes social knowledge extracted from user-generated content from Twitter to identify the social media users’ concerns on different domains of innovation regarding Sustainable Development Goals (SDGs) during 4 years, whereby understanding the emotional expression of public opinion upon those SDGs innovation dimensions. Topic analysis with latent semantic approach is the most suitable approach for exploring topics of interest from large text corpus; while sentiment analysis using Python-based library Vader is effective to investigate the Twitter users’ sentiment underneath those extracted topics. The importance of gender equality and youth empowerment on innovation; and the innovation in sustainable agriculture, education, eco-friendly materials, green energy, and economic development are intensively discussed topics. The component-terms of corresponding topics in different years are further scrutinized to highlight the evolution of Twitter users’ concerns on a particular innovation dimension. The result of sentiment analysis suggests the predominance of “very positive” and “positive” sentiment in almost topics during 4 years. The “neutral” sentiment prevails in a certain topic that generally addresses many angles of SDGs innovation in its majority of tweets without focusing on any specific dimensions. The high of “negative” sentiment during 4 years is noticed in the agriculture innovation topic in 2018.