Skip to content

Joy879/Sulify

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Sulify

A simple song recommender

sulify

About the app 💡

A song recommender that takes in data from Spotify API and displays various aspects in a dashboard. I wanted to explore how recommender systems work and use Python based tools for development since its the go to language for most Data Science projects.

  • View the webapp here

How it works 🐾

Sulify is a dash app. The structure is as follows:

- app.py
   |-- audio features
   |-- recommender

The audio features tab is meant to explore a single song's features. A user searches for a song and selects the specific song and artist from the dropdowm menu and gets to see a preview of the song, audio features displayed as numbers and displayed as graphs.

The recommender tab is meant to make requests to a model hosted as an API. The response is a list of ten songs that the user can preview and also view a graph of similarity scores.

Main libraries used: 📚

Tool/Library Purpose
Dash Dashboard design 😀
Dash Bootstrap Components Bring boostrap to dash apps
Plotly Interactive Graphs 📊
spotipy Connect to Spotify
scikit-learn Get similarity scores and graph 📈

Main Feature 📌

User searches for a song and either gets a summary of all audio features or a recommendation of similar songs all displayed in a visually pleasing manner.

All this is made possible through Dash. Dash has two main parts:

  • Layout - the HTML and DCC components that design the static webpage
  • Callbacks - Inputs and Outputs that define how the webpage becomes dynamic

Installation 📥

Prerequisites:

  • A Spotify free account which will give you access to an API key

            git clone https://github.com/Joy879/Sulify
            pip install -r requirements.txt
    

Licensing 🔒

This project is licensed under the MIT License - see the LICENSE file for details.

Future

I am looking forward to collecting more data to make the recommender model more accurate and faster. This project is my own way of exploring various strategies of implementing content-based recommender systems. The accuracy vs speed tradeoff is definitely an issue i'd like to dive deeper into even as I learn how to optimize the app to work with larger datasets. I would also like to allow a user to login to spotify from the app and be able to save the recommended songs into a playlist. If you have any ideas or suggestions or contributions feel free to reach out to me via mail 📧

Author ✒️

Joy Wanjiru

I am a data science enthusiast and a software engineering graduate from ALX and I love working with Python especially because of it's vast pool of libraries for scientific computing

Releases

No releases published

Packages

No packages published