Datasets:
license: mit
task_categories:
- text-generation
- text-classification
language:
- en
tags:
- music
- lyrics
- art
pretty_name: Genius Lyrics (English Subset)
size_categories:
- 100K<n<1M
source_datasets:
- brunokreiner/genius-lyrics
Genius Lyrics
This dataset is a processed, lightweight subset of the original brunokreiner/genius-lyrics dataset. It has been filtered to focus exclusively on English songs with valid artist data, making it optimized for NLP tasks involving English songwriting, lyric generation, or genre classification.
All original credit goes to Bruno Kreiner for scraping and compiling the original data. This version is simply a filtered downsize for ease of use.
Processing & Filtering
The following processing steps were applied to the original dataset to create this version:
Artist Validation: Removed all rows where the artist_name was missing or null.
Language Filtering: Filtered strictly for English songs (where is_english was marked as True).
-Feature Reduction: Dropped extraneous metadata columns to reduce file size and complexity. Only the essential columns for lyric analysis were retained.
Dataset Structure
The dataset consists of a CSV file containing the following columns:
artist_name (string): The name of the recording artist or band.
genres_list (string/list): A list of genres associated with the track.
lyrics (string): The full text of the song lyrics.
Usage
You can load this dataset directly using the Hugging Face datasets library:
from datasets import load_dataset
dataset = load_dataset("theelderemo/song-lyrics")
Credits & License Original Dataset: brunokreiner/genius-lyrics
Original Author: Bruno Kreiner
Please refer to the original dataset for specific licensing details regarding the scraped content.