Spotify_Audio_Analysis

A visual analytics system that enables users to understand what really makes a song popular through audio analysis. It supports high-level and low-level audio feature analysis. This repo mainly consists of Python scripts used to fetch and create our dataset, as there is no readily available dataset. Billboard Year-End Charts Top 100 songs from 2010 to 2023 are used to determine the most popular songs on a year-on-year basis. This repo mainly consists of scripts that were used for dataset creation. Another repo has the related visual analytics system that uses this data.

Scripts

• billboard_charts.py - Scrapes Billboard charts year-on-year from Wikipedia
• spotify_api_data.py - Main script that fetches data using Spotify API and scraped charts
  • get_spotify_tracks_from_playlist() - Uses Get Playlist Items API to get individual track IDs from playlists (that were manually searched/created)
  • get_spotify_audio_features() - Uses Get Audio Features API (deprecated - November 27, 2024) to get high-level audio features for each track using track ID

Data

• billboard_data.json - Scraped Data from Wikipedia that contains Billboard Year-End Top 100 Charts
• spotify_playlists.json - Manually curated playlist URLs that correspond to Billboard charts
• spotify_track_ids.json - Track IDs from playlists using get_spotify_tracks_from_playlist()
• track_features.json - Audio features for each track using get_spotify_audio_features()

Installation

1. Clone the repository
2. Create a virtual environment using python -m venv venv
3. Install packages using pip install -r requirements.txt

Using Spotify API

1. Create a web app in spotify dashboard. They have the necessary instructions on how to do so.
2. Create the following file env/secrets.py

class Spotify:    
    CLIENT_ID = '<client_id>'
    CLIENT_SECRET = '<client_secret>'
    ACCESS_TOKEN = '<access_token>'

3. Run generate_access_token() in spotify_api_data.py to get the access token once CLIENT_ID and CLIENT_SECRET are filled in