carlosfg97 / music-lyrics-analysis

An analysis of lyrics in rap music, involving web scraping, cosine similarity, MDS plots, LDA and topic modeling

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Multi-dimensional Lyrical Analysis

This projects explores the lyrics of several rap and hip-hop songs across several years to try and understand:

  • Lyrical similarity of prominent rap/hip-hop artists over the years, done using Cosine Similarity and Clustering
  • Musical influences of prominent rap/hip-hop artists over the years (i.e. if a given artist was inspired by another prior to his/her time based on lyrics), done using LDA and Jensen-Shannon distance.
  • General themes for which song lyrics can be classified, used to identify the 'topics' or 'subjects' that any given rapper seems to veer toward and/or for songs from a given time period.
  • Change in music trends throughout the years in terms of the themes/topics rappers in general seem to discuss in their music.

Dataset

The Spotify service was used for this project; several public playlists containing large volumes of hip-hop and rap music were curated and the SpotiPy Python API was used to retrieve track information.


Code contained here is used within the course 'INSY 669 Text Analytics' of the MMA Program at McGill University.

This project was done in collaboration with Carlos Fabbri, Matthieu J., Louis D'Hulst and Ananya Nair

About

An analysis of lyrics in rap music, involving web scraping, cosine similarity, MDS plots, LDA and topic modeling


Languages

Language:Jupyter Notebook 100.0%