Inclusion in Beauty: A Data-Driven Approach to Measuring Diversity in Foundation Shades

Presenter: Christina Pathrose

Research Category: Computer and Information Sciences
College: Khoury College of Computer Sciences
Major(s): Computer Science
Student Type: Undergraduate
Graduation Date: 2022
Additional Authors: Gaea Leemon
Award Winner Category: Computer and Information Sciences

As a self-expressive art form, makeup is no longer viewed as a womanÕs way of obscuring her insecurities. With the rise of social media, makeup has evolved into a medium of communication in which artists Ñmale, female and non-binary – can depict stories with each brush stroke, share their confidence, inspire others, and as seen with many movements, it even has the potential to spark change. Most looks begin with foundation, which is a powder or liquid that is applied to the face in order to level or even oneÕs complexion. Yet, for many people of color, the process of finding the right shade of foundation can be discouraging. As of 2020, the cosmetic industry was estimated to be valued at 49.2 billion USD, in which people of color contributed an estimated two-thirds of revenue. However, most foundation aisles across America consist primarily of lighter foundation shades with the exception of a few dark shades that fail to cater to the diverse range of skin colors. This project seeks to use machine learning to better understand how inclusive the beauty industry is today for people all across the diverse spectrum of skin colors. Using a dataset published by The Pudding as our baseline data, we shall add corresponding price and accessibility data. We intend to measure the diversity of shades available, price variance for different shades, accessibility of shades, as well as train a machine learning algorithm to recommend foundation products to a user given their complexion.