Name: 40K Songs With Audio Features And Lyrics
Creator: Kaggle
License: http://opendatacommons.org/licenses/odbl/1.0/

About this Dataset

40K Songs With Audio Features And Lyrics

The 40K Songs with Audio Features and Lyrics dataset contains about 41000 songs along with their full lyrics and explicit audio features, making it a versatile source for developers or students working on AI applications in the music domain, such as music recommendation systems, building classification models, or even running sentiment and emotional analysis.

It was created by carefully merging and preprocessing the following three datasets:

Spotify_1Million_Tracks By: Amitansh joshi, provided detailed audio features.
Song Lyrics from 79 musical genres By: Anderson Neisse, offered a wide collection of song lyrics across a variety of gernes.
Grunge Bands Dataset By: Ana Vučić

With Spotify restricting access to several previously available API endpoints (including the ones used to retrieve audio features), many individuals can no longer fetch the required information for their projects. This dataset fills that gap by providing a ready-to-use collection of songs lyrics and audio features, enabling experimentation and model development without using any API.

Don't forget to upvote if you find this dataset useful! And please if you have feedback or any idea for expansion let me know in the discussion section :D

Tables

Final Dataset V4

@kaggle.zinasakr_40k_songs_with_audio_features_and_lyrics.final_dataset_v4

26.02 MB
41,574 rows
19 columns

CREATE TABLE final_dataset_v4 (
  "artist_name" VARCHAR,
  "song_name" VARCHAR,
  "language" VARCHAR,
  "lyrics" VARCHAR,
  "genres" VARCHAR,
  "new_artist_popularity" DOUBLE,
  "acousticness" DOUBLE,
  "danceability" DOUBLE,
  "energy" DOUBLE,
  "instrumentalness" DOUBLE,
  "key" DOUBLE,
  "liveness" DOUBLE,
  "loudness" DOUBLE,
  "mode" DOUBLE,
  "speechiness" DOUBLE,
  "tempo" DOUBLE,
  "time_signature" DOUBLE,
  "valence" DOUBLE,
  "duration_ms" DOUBLE
);