From Words to Action: Sentiment Analysis on Sustainability Initiatives

Sentiment analysis is pivotal in gauging public perceptions and attitudes toward sustainability initiatives. Understanding how individuals express sentiments related to sustainability on social media platforms is crucial for assessing the effectiveness of these initiatives and tailoring communication strategies. This paper presents a comprehensive sentiment analysis study focusing on sustainability-related tweets. Leveraging a combination of pre-trained models, including VADER, TextBlob, and Flair, and traditional machine learning classifiers such as Logistic Regression, SVM, KNN, Decision Tree, Random Forest, and Naive Bayes, we explore diverse methodologies for capturing sentiments in the context of sustainability. The experimental setup involves three distinct datasets with keyword-based filtration, showcasing the challenges and nuances within sustainability sentiment analysis. Performance metrics, including precision, recall, F1-score, and accuracy, are reported for each model and dataset, revealing the strengths and limitations of various approaches. The combined model comparison across datasets provides insights into the overall efficacy of sentiment analysis methodologies. In conclusion, this study opens avenues for future research in sentiment analysis within the realm of sustainability discourse. By shedding light on the complexities of sentiments expressed in sustainability-related tweets, our research lays the groundwork for refining existing models and exploring innovative approaches.