Analysing sustainable finance cluster on Twitter with topic modelling and sentiment analysis

Day by day, the global climate crisis threatens all the people on the world. The climate crisis comes along with many environmental (rising sea levels, drought, heat waves, etc.) and social risks (migration, income injustice, etc.). The financial sector, on the other hand, bears a significant amount of responsibility for ensuring sustainable development and providing the capital transfer that developing countries require in order to combat and adapt to climate change and its social and environmental consequences. Considering these factors, sustainable finance has the great potential to create a more livable world for those living on mother earth and above it. Firstly, the conceptual framework of sustainable finance is mentioned. Considering that there are very few studies on sustainable finance and ESG investments in Turkish publications, the conceptual framework is very important. In the second part of the study, an English database is created by searching for keywords related to sustainable finance on Twitter with the help of Twint Library. The database was interpreted by data visualizations. After the data is preprocessed, the LDA and LDAmulticore models called from the Gensim library and the LDA mallet technique called from the Mallet library are compared by topic modeling and the optimal number of subjects is tested. Sentiment analysis was performed on the modeled topics and tweets using the textblob library. As a result of our application, 18 topics in LDA mallet model that give the highest coherence score are determined, and these topics are listed with keywords. The trends within the sustainable finance community are analyzed comparatively according to the topics and their sentiment analysis. The study is the first known example of topic modeling in the field of sustainable finance.