Enhancing Sustainability Reporting Through Automated Text Analysis: A Systematic Review and Empirical Study in the Apparel Industry

Abstract

This thesis investigates the transformative potential of automated text analysis techniques, specifically text mining (TM) and natural language processing (NLP), in advancing the analysis of sustainability reporting (SR). Through two interrelated studies, critical gaps in the field are addressed by evaluating the current application of text analysis to SR analysis and introducing a novel analytical framework. The first study offers a comprehensive systematic literature review (SLR) of TM and NLP applications in SR, assessing methodologies, research objectives, and analytical depth. The findings underscore both the promise and limitations of automated text analysis in extracting meaningful insights from SRs, highlighting untapped potential. Building on these insights, the second study applies advanced text mining techniques to a case study within the apparel industry, focusing on the years following the Rana Plaza disaster. This application reveals biases toward positive reporting, raising concerns about the transparency and credibility of SRs. Together, these studies enhance the theoretical understanding and practical application of TM-NLP tools in SR, advocating for more transparent, balanced, and credible sustainability practices.