mwdchang / songvis

Song attributes and lyrics visualizations.

songvis

A visualization of your team's playlist, leveraging Spotify's track attributes and other analytics.

Getting data

Requirements

We need the following to run the parsers and the web application

• We need a Spotify account and API-access key to retrieve song attributes, see https://developer.spotify.com/documentation/web-api/quick-start/
• NodeJS 12+
• Python 3.7/3.8

Input file

A CSV formatted file with 3-coumns, tracking the person, song/track id, and the lyrics. Note the lyrics are surrounded by speical pipe(|) character in order to mark start/end boundaries. See provided "sample.csv".

<name a>,<trackId>,|<lyrics>|
<name b>,<trackId>,|<lyrics>|
<name c>,<trackId>,|<lyrics>|
...
<name a>,<trackId>,|<lyrics>|

Track analysis

Track data are fetched via Spotify's API.

cd parsers

# Rename sample.env to .env and fill out the access credentials

npm install
node extracter.js <input.csv> > tracks.tsv

Lyrics analysis

Lyrical similarity and others. Using GenSim

cd parsers
pip3 install -r requirements.txt
python3 vectorize.py <input.csv> > lyrics.tsv

Running

Running the application. Place the fiels "tracks.tsv" and "lyrics.tsv" into the public folder.

npm install
npm run serve

The application will be available on http://localhost:8000, it will take a few seonds to load and process the data.