I built a recommendation model for Skincare Products.

Beauty technology and other endeavors

tldr; I built a recommendation model for skin care products using scraped reviews from Sephora’s website. It’s available as a web app due to integrating data tool streamlit.

Here it is on my Github!

Long version

15 months ago I decided to teach myself Python. This had been something I put on hold for nearly a decade as I pursued a music career in Tokyo. Not much time for coding.

But last year I had the perfect opportunity — I left my b2b SAAS job and had the resolve to as learn as much as possible before I began my graduate school program.

After a month of learning I started to get momentum.

I had the most fun scraping data and making prediction models — I really enjoyed doing one for MyAnimeList and Sephora.

This came to a pause early Summer 2023. I hit a major set back in my personal life. I went through my first serious break up & beginning an exciting new start up job in a matter of weeks.

Not much time for doing projects. My day to day work at Mintstars allows me to do technical tasks however it’s primarily in the form of data analysis.

So I’m trying something new - to finish each project that I start. Monday this week I decided I would go back to the Kaggle dataset I played around with last year and would create a working model from start to finish, deploy it to a web app and commit it to Github!

My goal is to ship at least one project a week.

As for this project — There’s a lot of work to be done on the UI side of things and making the model more accurate.

I’ve considered having it be a basic dupe machine that allows people to find cheaper alternatives to popular skin care items based on their needs.

I’m working on a larger project at the moment and am quite busy with work so this will have to be something that I come back to someday if I feel especially inspired. If you have any suggestions, ideas or open source projects you think I’d enjoy please send them my way :)