Using NLP to ascertain changes in the fast-fashion industry based on UN sustainable development goals

(1) Westview High School, Portland, Oregon, (2) Department of Computer Science, Portland State University, Portland, Oregon

* These authors made equal contributions

https://doi.org/10.59720/22-214
Cover photo for Using NLP to ascertain changes in the fast-fashion industry based on UN sustainable development goals
Image credit: Prudence Earl

The fast fashion industry largely contributes to pollution and global emissions, which can be mitigated through strengthening policy. Policy change requires knowledge of the changing conditions of the fast fashion industry that can be evaluated through the sustainability reports published by fast fashion companies. However, due to the reports’ variations in details and lengths, it is difficult to find trends that promote the United Nations Sustainable Development Goals (UN SDGs), such as “Responsible Consumption and Production” and “Climate Action”. Therefore, we ask if fast fashion companies have changed their sustainability focuses in terms of the UN SDGs from 2018–2019 to 2020–2021. We hypothesized that the later years would showcase higher similarities towards the UN health-related goals due to COVID-19. After checking the reports’ similarities against the keywords of each UN SDG, we found that the similarity percentage of the “Good Health and Well-Being” goal had more than doubled from 2018–2019 to 2020–2021. This is important because it shows how the industry reacts to global changes like a pandemic. We also noticed that the three most frequently addressed goals were the same for both time periods. Specifically, the most dominant topic was “Responsible Consumption and Production,” with a similarity score of 15.54% for 2018–2019 and 14.1% for 2020–2021. This suggests that the industry may have some consistent focuses on sustainability over time. With these insights, policymakers can better address what changes the fast fashion industry needs to make.

Download Full Article as PDF

This article has been tagged with: