Recommendation System: Data Preprocessing

Cleaning the data and preparing it for training

  • Bijon Setyawan Raya

    • October 25, 2023

    2 mins

    Music Recommendation Sytem (4 Parts)

    Data Collection

    Data Preprocessing

    System Training

    System Evaluation

    Introduction#

    I am going to seperate this post into two parts: the technical part and the insight part. The technical part consists of how I prepared, passed, and predicted the data. While the insight part consists of the result of the prediction and what the prediction tells me.

    Music Features / Aspects#

    The dataset consists of my favorite songs from my own playlists, and each column represents different aspects of the song.

    Each aspect represents:

    1. popularity
    2. acousticness
    3. danceability
    4. duration_ms
    5. energy
    6. instrumentalness
    7. key
    8. liveliness
    9. mode
    10. speechiness
    11. tempo
    12. time_signature
    13. valence

    For more details what each aspect entails, you can check out Spotify's documentation.

    Technical Part#

    Insights#