AI Generated Lyrics from Your Favorite Beatle

Contents

• Notebooks
  
  Three separate notebooks describing the process of getting the data, analyzing the data, and creating the lyric generator the resulting model
• /Data
  
  The various files containing the data I collected and used
• /helper_functions
  
  Python files that make the notebooks work
• /Images
  
  Important images used in the analysis and presentation
• streamlit_app.py
  
  Run this file to explore the lyric generator app
• Presentation Slide Deck
  

Description

This projects explores the lyrics behind The Beatles through NLP with NLTK and NMF. Also, there is a lyric generator based on The Beatles using GPT-2 and streamlit.

Tools Used

• Natural Language Processing
• Topic Modeling
• Sentiment Analysis
• NLTK
• NMF
• GPT-2
• Seaborn
• Matplotlib
• Numpy
• Pandas
• JSON

Potential Impact & Use

The lyric generator itself is mostly for fun, but the concept of a lyric generator can be used by artists. As a songwriter, I plan on creating my own lyric generator trained on lyrics I’ve written to help me out when I’m in a rut. As for the analysis, one specific insight popped out, and that was the fact that Ringo as the most negative sentiment on Rubber Soul. I always thought Ringo sang about funny topics, but this proved my assumption wrong.