ANALYZING CORPORATE SUSTAINABILITY REPORTS WITH TEXT ANALYSIS METHOD: EVIDENCE FROM BIST SUSTAINABILITY 25 INDEX

Abstract

Sustainability reports enable businesses to inform the public about their progress towards their
goals, including environmental, social and management measures, and the risks they may face now
or in the future. Because businesses play an active role in the sustainable development process,
investors are more interested in the activities of businesses and the impact of these activities on the
environment. In this study, the sustainability reports of the companies included in the Borsa
Istanbul (BIST) sustainability 25 index were evaluated. Using a sample of 16 companies included in
the Borsa Istanbul sustainability 25 index and publishing sustainability reports in 2021, text analysis
was conducted to identify trends in sustainability reports. The analysis process was carried out
using the Python programming language. According to the results of the analysis obtained, it was
determined that 81% of the sustainability report statements showed positive sensitivity and 19% had
negative sensitivity. In addition, in the study, LDA topic modeling and distance mapping were
applied to 16 sustainability reports, and it was observed that reports 2, 15, 6, 13 and 8 overlapped
with each other as a result of the application. It was assumed that there was no definitive word list
for each report, as no heavy overlap was found between the other reports. Therefore, we can state
that the PyLDAvis package included in the Python software is an effective tool for identifying
overlaps and the most important themes among sustainability reports.