View Code? Open in Web Editor
NEW
Home Page: https://songrecommendation.anandgmurugan.me
Jupyter Notebook 89.61%
Python 10.39%
song-recommendation-qdrant's Introduction
Song-Recommendation-Qdrant
Clone this repo into a directory of your choice.
Install Qdrant via Docker. (refer:https://qdrant.tech/documentation/quick-start/ )
Note: The commands are shown in UNIX-shell, refer qdrant_commands_for_windows.txt
for the windows alternatives.
The ipynb notebook is for demonstration purposes.
For the actual app, navigate to the directory in your terminal and run streamlit run song_rec_streamlit.py
.
This is an implementation of a recommendation engine using the Qdrant Recommendation API.
We source the data from https://www.kaggle.com/datasets/geomack/spotifyclassification . (Note: You can get your own data using the spotify API).
The data is vectorized and stored in a Qdrant collection.
The Recommendation API takes positive and negative feedback from the user as well as artist filters and uses vector similarity to assign a score to each of the songs in our data.
The top 5 songs are displayed back to the user.
song-recommendation-qdrant's People
Contributors
Stargazers
Watchers