Santos Discography / Lyrics Analyser

Ever wondered if those positively charged songs are more successful than others? How has the sentiment of a musician’s lyrics changed over time? How rich is their vocabulary or how popular are their individual albums among fans? Then you’re in the right place.

Santos Discography / Lyrics Analyser is a Jupyter Notebook (Python) based interface to download and visualise music artists’ discography data, statistics and lyrics sentiment.

High level overview:

For any given musician/band name:

• retrieves the discography of studio albums,
  
• allows user to select albums from a list, then retrieves lyrics for songs on selected albums,
  
• collects data such as owners and album ratings from DISCOGS.COM, album and song placement in official Billboard music charts,
  
• performs lyrics sentiment analysis and generates scores,
  
• with a simple user interface, allows users to browse a number of visalisations (by user defined period intervals where applicable).

Quick demo

Example visualisations:

• number of songs and albums over time [Example: David Bowie]

• lexical diversity over time [Example: Eminem]

• word clouds by album or period [Example: David Bowie]

• Discogs album ratings and statistics [Example: Depeche Mode]

• lyrics sentiment scores across artist’s discography [Example: Eminem]

• lyrics sentiment scores by song in specific album [Example: Eminem]

• success of tracks and albums in Billboard charts over time

• success in Billboard charts vs lyrics sentiment [Example: David Bowie]

• lyrics sentiment over time [Example: David Bowie]

Dependencies:

APIs and web scraping

lyricsgenius
lxml.html
cssselect
discogs_client
requests
lyricsmaster

data processing and charts

matplotlib
pandas
numpy
time
seaborn
fuzzy_pandas
chord

UI

ipywidgets
IPython
tqdm

word processing

re
os
nltk
string
wordcloud