New App: NLP analysis of coffee reviews for recommender

This is my first app on Streamlit that makes recommendations (and a little more) for coffee based on topic modeling of online coffee reviews. I’d love any feedback for Streamlit, Python, or if there are any issues:
https://share.streamlit.io/ejfeldman7/coffee-reviews-nlp/main/web_app/web_app.py

1 Like

Awesome!

For those of us who don’t know how to describe our coffee :laughing:, it would be great to have an “I’m feeling lucky” button where you randomly pick an example description from a set.

Also, I’m getting a the error below when I enter a description:

2 Likes

That’s a great idea, thanks, and appreciate helping me find items to debug.

Nice App - I like playing with different terms on the ‘Recommender’ page. Did you consider using a radio button there so you can easily switch between the two options?

That’s a good thought for a change, thanks. I had intended to go back and fix that page, it’s a little clunky, and I think a radio button would definitely be nicer.