Sentimental Analysis of Environmental Discourse on Twitter: A Computational Corpus Based Study

Abstract

This paper examines the level of sentimental orientation and subjectivity integrated within environmental discourse on Twitter (X). Based on a self-assembled collection of 100 purposely sampled tweets on environmental topics specifically those about Pakistan climate crisis. TextBlob, a Python natural language processing library, is used to derive polarity and subjectivity rating of each tweet. Theoretical background relies on the Appraisal Theory (Martin and White, 2005), the rate of textual evaluation  and Sentiment Polarity Theory which categorizes textual impact on a scale between positive and negative. The results show that positive sentiment is in 40 percent of the tweets, negative 38 percent and neutral 22 percent, which provide on average a positivity of +0.0011, which is a slight positive, but still positive, balance of the environmental conversation. The negative sentiment is primarily related to climate disasters, pollution, and failure in governance whereas positive sentiment is observed in the conversation around renewable energy, community resilience, and climate activism. The results also demonstrate that evaluative adjectives, intensifiers and emotionally colored words are significant to influence sentiment polarity. On the whole, this paper has proven that the discourse on environmental issues on social media is extremely critical and contains the elements of the crisis-related discourses and those of the solution-oriented ones.