Introduction
Don't know where it was produced? Don't know where it came from? All you need
to know is how your feel about it and your description. This app will list
all possible wines' information. Perhaps there is one you are looking for.
Motivation & Purpose
Looking for wines you like is like picking needle from ocean. Looking through
bunch of reviews is a pain or sometime your have no clues where to start.
Use your sentiment about a wine for searching could shorten this process.
More
The machine learning model was trained with dataset from Kaggle.
There are 2 models, old model and new model. And they use different training method.
Old model: supervise learning
New model: unsupervise learning
Both models find the similarity of description to predouce final result.
Frameworks were used to build this app:
🧰 Scikit-Learn : Machine learning framework for model
🧰 Streamlit : Backend with frontend app for new version
🧰 Nextjs : Frontend app for old version
🧰 Docker : Containerized backend for old version
Note
Old version only provide wine's title, country, score, price(💲USD) and description.
Backend server often put to sleep if no interaction for period of time. Therefore
website(frontend) will take a while to wait for server to wake up to be able to function.
New version only provide wine's location information.
Website