classes.py

from spotipy import oauth2
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import requests

from secrets import client_id, client_secret


class ArtistAnalysis:
    def __init__(self, artist_id, base_url,):
        # AUTHORIZATION
        self.auth = oauth2.SpotifyClientCredentials(
            client_id=client_id,
            client_secret=client_secret
        )

        self.auth_url = 'https://accounts.spotify.com/api/token'
        # POST REQUEST
        self.auth_response = requests.post(self.auth_url, {
            'grant_type': 'client_credentials',
            'client_id': client_id,
            'client_secret': client_secret,
        })
        # CONVERT RESPONSE TO JSON
        self.auth_response_data = self.auth_response.json()
        # SAVE ACCESS TOKEN
        self.access_token = self.auth_response_data['access_token']

        self.headers = {
            'Authorization': 'Bearer {token}'.format(token=self.access_token)
        }

        self.base_url = base_url
        self.artist_id = artist_id

    def get_artist_tracks(self):
        # GET REQUEST
        r = requests.get(self.base_url + 'artists/' + self.artist_id + '/albums',
                         headers=self.headers,
                         params={'include_groups': 'album', 'limit': 50})

        d = r.json()

        data = []  # HOLDS TRACK INFO
        albums = []  # HOLDS ALBUMS (NO DUPLICATES)

        # LOOP OVER ALBUMS FOR TRACKS
        for album in d['items']:
            album_name = album['name']
            print(album_name)

            # GET TRACKS FROM ALBUM
            r = requests.get(self.base_url + 'albums/' + album['id'] + '/tracks',
                             headers=self.headers)
            tracks = r.json()['items']

            for track in tracks:
                # GET AUDIO FEATURES (KEY, DANCEABILITY, ETC.)
                f = requests.get(self.base_url + 'audio-features/' + track['id'],
                                 headers=self.headers)
                f = f.json()

                # COLLECT ALL ALBUM INFO
                f.update({
                    'track_name': track['name'],
                    'album_name': album_name,
                    'release_date': album['release_date'],
                    'album_id': album['id']
                })

                data.append(f)
        return data

    def visualize_audio(self):
        # NOW TIME FOR DATAFRAMES
        track_data = self.get_artist_tracks()
        df = pd.DataFrame(track_data)
        # CONVERT RELEASE DATE TO ACTUAL DATE AND SORT
        df['release_date'] = pd.to_datetime(df['release_date'])
        df = df.sort_values(by='release_date')
        df.head()

        plt.figure(figsize=(10, 10))

        # BUILD SCATTER PLOT
        ax = sns.scatterplot(data=df, x='instrumentalness', y='speechiness',
                             hue='album_name', palette='rainbow',
                             size='duration_ms', sizes=(50, 1000),
                             alpha=0.7)

        # DISPLAY LEGEND WITH NO SIZE ATTRIBUTE
        h, labs = ax.get_legend_handles_labels()
        ax.legend(h[1:10], labs[1:10], loc='best', title=None)
        ax.legend()
        ax.grid(True)
        plt.show()
        return


if __name__ == '__main__':
    base_url = 'https://api.spotify.com/v1/'
    artist_id = "1a4N2lwra7WGjwCDJS1Dkk"
    Lotus = ArtistAnalysis(artist_id, base_url)
    Lotus.visualize_audio()