When I first began working on this project I wanted to see if there were any commonalities between my favorite artists and songs within k-pop. The scope slowly grew bigger and bigger over the months. Now, not only has data been scraped and a recommender system built, but a micro framework using flask has also been constructed for users to interact with the recommender engine. Future goals for this project include hosting everything in the cloud and scheduling regular scraping of Spotify to ensure data contained within the recommender engine is up to date.

In order to run the “website” on your end, see the following steps:

1. First clone the repo to a desired location: git clone https://github.com/ericwan1/kmusic-recos.git
2. Open terminal or command prompt and navigate to the repository.
3. Run export FLASK_APP=app or set FLASK_APP=app if you are using the latter.
4. Then run flask run.
5. Following the output from step 4, you will notice that the framework is most likely running on the default port (5000). Follow the url provided http://127.0.0.1:5000 and you can begin playing around with the site!

• The model uses nearest neighbors. Scraping was done using the Spotipy API, and a basic analysis notebook is also included.
• Flask was selected due to ease of set up, features, and easy integration with the code I already wrote.

• 273_kArtists.csv contains 273 artists and their Spotify IDs
  • Artist (Artist Name)
  • Spotify_Id (Artists’ Unique Spotify ID)
• album_to_track_df.csv houses artist, album, and track information
  • Artist (Artist Name)
  • Artist_Id (Artists’ Unique Spotify ID)
  • Album_Name (Album Name)
  • Album_Id (Album’s Unique Spotify ID)
  • Track_Title (Title of the Track)
  • Track_Id (A Track’s Unique ID)
  • misc. song information – see bottom of dictionary
• artist_appears_on_df.csv contains information on other albums the artist appears on
  • Artist (Artist Name)
  • Artist_Id (Artists’ Unique Spotify ID)
  • Appears_On_Name (“Albums” the Artist has Appeared On)
  • Appears_On_Id (The ID of the “Album” the Artist Appeared On)
• compilation_track.csv includes artist information and compilation information, as well as track information for tracks contained within the artist compilation
  • Artist (Artist Name)
  • Artist_Id (Artists’ Unique Spotify ID)
  • Compilation_Name (Compilation Name)
  • Compilation_Id (The Compilation’s Unique Spotify ID)
  • Track_Title (Title of the Track)
  • Track_Id (A Track’s Unique ID)
  • misc. song information – see bottom of dictionary
• single_album_track_data.csv contains artist information, any “single albums” and tracks contained within said albums and track information
  • Artist (Artist Name)
  • Artist_Id (Artists’ Unique Spotify ID)
  • Single_Album_Name (Name of the Single Album)
  • Single_Album_Id (Artists’ Unique Spotify ID)
  • Track_Title (Title of the Track)
  • Track_Id (A Track’s Unique ID)
  • misc. song information – see bottom of dictionary
• misc. song information
  • danceability (Track Suitability to Dancing, 0.0 to 1.0)
  • energy (Measurement of Intensity and Activity, 0.0. to 1.0)
  • key (The Key of the Track – see pitch class table Guide Here)
  • loudness (Overall Loudness in Decibels, Typically -60 to 0 db)
  • mode (Major or Minor, 1 or 0)
  • speechiness (Presence of Words – the higher, the more wordy. 0.0 to 1.0)
  • acousticness (Confidence Measurement of a Track Being Acoustic. 0.0 to 1.0)
  • instrumentalness (Predicts if a Track Contains no Vocals, 0.0 to 1.0)
  • liveness (Predicts if a Live Audience in the Track, 0.0 to 1.0)
  • valence (Musical Positiveness, the Higher, the ‘Happier’. 0.0 to 1.0)
  • tempo (Track’s Estimated BPM)
  • duration_ms (Track Duration in Milliseconds)
  • time_signature (Time Signature, 3 to 7 for 3/4 to 7/4)

• Host everything on a server and make the website interactable to the world
• Automate the scraping notebook and orchestrate runs; potentially store and run everything in the cloud