Building an “ethical-rating” classifier

This is the first post in a series where I will be documenting the progress of an ethical rating classifier as part of a personal project in collaboration with Mei Li and Chih-Te Courtney Hsu.

Background

The year is 2021 and consumers are becoming acutely aware of unethical labor and sustainability practices within the fashion industry. From spending boycotts, to built-in plugins blocking unsustainable brands, consumers are getting to combat companies with unethical practices.

But so are fashion companies.

Greenwashing. A process in which companies convey a false impression or provide misleading information about how a company's products are more environmentally sound.

Many fashion companies draw attempts to establish ethical reputations by releasing statements on issues such as environmentalism. Despite the strong language, however, consumers are often skeptical of the messaging. Websites, such as Good on You, attempt to educate consumers of company practices; often identifying “greenwashing” statements through their platform. But educating oneself is a time-consuming process, so how could consumers become more aware of these so-called “green-washed” statements?

Adopting similar methods for building spam detection models, we approached this problem from a data science perspective.

In this project, we will explore the following questions:

1. What aspects of online brand statements predict company ethical ratings, specifically in the athleisure industry?

2. What is the impact of sustainability practices of a brand for customer fashion purchases?

With these questions in mind, our proposed solutions include:

• Building a classification model to identify ethical and unethical companies based on language used in company statements

• Assessing consumer perceptions on topics of company ethics and sustainability practices based on fashion subreddit posts and comments

Curious to see our progress? Check out our repo using the link below:

This page will be updated as we proceed with our project.