Sustainability reporting, particularly through Corporate Social Responsibility (CSR) disclosures, plays a critical role in assessing corporate contributions to environmental and social goals. While much attention has been given to other environmental dimensions in CSR reports, waste management remains an underexplored yet significant area. This research addresses this gap by developing methodologies to analyze and classify waste management disclosures, distinguishing between substantive and symbolic CSR reporting.

A key challenge in CSR reporting is the presence of boilerplate language and greenwashing, where companies present sustainability commitments without substantive action. This study differentiates symbolic CSR, which prioritizes corporate image over tangible sustainability efforts, from substantive CSR, which reflects genuine, measurable contributions to waste management. Using a pre-processed dataset of 9,500 CSR reports across multiple industries, the research employs natural language processing (NLP) techniques, including Stanford's Stanza NLP model, to automate the classification of CSR disclosures. The methodology involves sentence segmentation, dependency parsing, and keyword-based classification, ensuring a robust distinction between symbolic and substantive statements.

Furthermore, this research examines how waste management disclosures vary across different industry sectors and evaluates their alignment with the Sustainable Development Goals (SDGs). By integrating computational text analysis with manual validation, the study enhances the methodological rigor in assessing CSR transparency. The findings contribute to both theoretical and practical discussions on CSR reporting, offering insights for policymakers, investors, and sustainability advocates seeking to hold corporations accountable for their environmental commitments.