Green Finance in the Digital Public Sphere: A Multi-Method NLP Analysis

Green finance has become a key component of sustainability transitions by supporting the alignment of financial systems with environmental and climate-related objectives. As discussions on sustainable investments and climate finance increasingly take place in digital environments, analyzing large-scale user-generated content has become important for understanding online discussions and emerging discourse patterns. This study investigates green finance discourse in the digital public sphere using a multi-method Natural Language Processing (NLP) framework. A dataset of 23,452 posts collected from the platform X was analyzed using TF-IDF-based word frequency analysis, FinBERT-based sentiment analysis, Latent Dirichlet Allocation (LDA) topic modelling, and keyword co-occurrence network analysis. The results indicate that green finance discourse is dominated by positive (44.61%) and neutral (45.85%) sentiment, suggesting a generally favorable and institutionalized public perception. Topic modelling identified three dominant themes: climate finance and sustainability transition, digital finance and speculative investment narratives, energy investments and financial infrastructure. Network analysis revealed that sustainability, climate, energy, investment, and finance constitute the core conceptual structure of discourse. These findings demonstrate the effectiveness of NLP-based approaches for analyzing large-scale digital discussions and provide insights for policymakers, financial institutions, and organizations seeking to better understand online discussions surrounding green finance and support more effective sustainability communication strategies.