Skip to content

README.md

Latest commit

 

History

History
94 lines (63 loc) · 2.79 KB

README.md

File metadata and controls

94 lines (63 loc) · 2.79 KB

Comparify

Compare artists songs from Spotify using the sound features available in the spotify official API.

Installation

  1. clone the repo

  2. go to repo folder and type:

pip install -r requirements.txt
python setup.py install

Tutorial

Comporify under de hood uses the official Spotify API (spotipy) so you will need to create an authorized app token in Spotify Developer website: https://developer.spotify.com/

Then create a credentials manager with your cliend_id and client_secret:

from spotipy.oauth2 import SpotifyClientCredentials

client_id = '<your_client_token>'
client_secret = '<your_client_secret>'
credentials = SpotifyClientCredentials(client_id, client_secret)

Then create go to spotify and find out your artist of choice url. For example my band has the url https://open.spotify.com/artist/3rD7bBI9zkYhu62o79tWe6

The artist_id is the last characters of the url: 3rD7bBI9zkYhu62o79tWe6

Then you can start playing around with it simply doing:

from comparify import SongFeatures

# one artist
dasouza = SongFeatures(credentials, '3rD7bBI9zkYhu62o79tWe6')
# another artist
malkmus = SongFeatures(credentials, '7wyRA7deGRxozTyBc6QXPe')

This returns a SongFeatures objects which basically is a pandas.DataFrame with extra information. It has the attributes of a DataFrame:

dasouza.head(5)
id name album loudness energy valence danceability tempo speechiness instrumentalness acousticness liveness
4aaco211p3HFubmFfBoaxj Noves Venècies Futbol d'Avantguarda -8.716 0.356 0.608 0.582 130.898 0.0321 0.126 0.452 0.118
4N5E7jIK4Ti9yXoshshQ1J Migracions de salmons Futbol d'Avantguarda -7.193 0.61 0.657 0.535 125.95 0.0299 0.255 0.134 0.131
0hdwI9Xf7GnJUyUmXNETt0 Finals Futbol d'Avantguarda -6.603 0.61 0.72 0.638 130.071 0.0304 0.0491 0.0598 0.14 0
ERzbcM3uYCmH0zuONy02y Tan enfora Futbol d'Avantguarda -6.709 0.723 0.537 0.607 120.025 0.0285 0.162 0.216 0.228
2IjcG82yOERIFe9bV4dWrL Dos microbis Futbol d'Avantguarda -9.703 0.422 0.384 0.375 153.432 0.0424 0.525 0.622 0.0983

Then instance the Comparator:

from comparify import Comparator
comparator = Comparator(scaler='minmax', projector='tsne')
comparator.fit([dasouza, malkmus])

And you will get a plotly scatter with the projected features in a 2d space.

Also you can get a the most similar song (euclidian distance based) with most_similar_song()

from comparify import most_similar_song

most_similar_song(dasouza, malkmus)
name closest
Noves Venècies Senator
Migracions de salmons Tigers
Finals Senator
Tan enfora 1% Of One
Dos microbis Us

Disclaimer

This is a toy project so feel free to say me anything or sue me because of my bad abstractions.