At idbeauty, we provide the tools for makeup enthusiasts of all skill levels to discover new products that work for their unique beauty.
Our project is a tool for makeup enthusiasts to discover new products and styles that work for their unique beauty, using the Co:here API LLM to classify text relating to makeup styles.
Find our submission on Devpost here.
This idea came about because of a conversation between us on our way to the hackathon! Getting into makeup as a minorities, we've all shared the experience of not knowing how to style our makeup to complement our features. There's a lot of information shared on social medias and by brands, and this can be overwhelming and confusing. We wanted to make a web application that can take our users concerns and suggest styles and products customized to them!
Using the style quiz and user text input, our platform provides suggestions on products and makeup styles that would best suit their desired look. We want to target all makeup enthusiasts, especially beginners, who wish to learn more about makeup techniques and products that match their goals. We hope to make information more accessible for all -- making learning about makeup less intimidating, more inclusive, and more fun!
We built idbeauty as a webapp using Node.js and React, with our design process mapped out on Figma. Our platform uses the Co:here API large language model (LLM) to analyze and classify text relating to makeup styles.
We first built our site using basic HTML and CSS, but came into issues when trying to implement the Co:here API. We then had to switch our entire project into Node.js/React...
This whole thing was a learning experience for us all! Two of our members come from biology/bioinformatic backgrounds, and one is a beginner in web design. We learned a lot of syntax, issue tracking, and design -- and best of all, had great teamwork and communication!
We learned
- HTML and CSS
- Node.js and React
- Figma
- The creative process in terms of developing a new product
- All work no sleep* (2 hours of sleep actually)
- Implementing reviews and further userbase communications
- Safeguards for reviews to avoid "review-bombing" (aka removing duplicate user reviews that skew results)
- Collection of more data and column categories to provide even more detailed style guides for more complex factors like acne-prone skin