Chordtext

• 

  The website is simple, but there's a homemade machine learning model behind it.

• 

  Localhost testing before deploying to PythonAnywhere.

--> DEMO VIDEO PASS: MLHROCKS

Do you have catchy lyrics for a song, but no clue what chord progressions to use?

Chordtext is a web app that uses sentiment analysis to determine whether a user-input set of lyrics would best fit a major or minor chord progression.

Below you'll find some (brief) takeaways from this spontaneous weekend solo hack. For a more technical run-down of how this works, check out Chordtext on GitHub!

The Fun Parts

• learning about & using sentiment analysis for the first time
• deciding to train an ML model on a Friday night
• remembering enough Python, HTML, CSS to host on PythonAnywhere
• brainstorming how to translate music theory concepts into code

The Slightly-Less-Fun Parts

• all your [Python dependency errors] are belong to me
• having to stop mid-hack to take a Multivariable Calc midterm
• realizing I couldn't use my domain with PythonAnywhere free

Though I didn't have as much time to hack this weekend as I would've liked, I had fun diving into a new subject area and immediately applying this new knowledge. In the future, I'd love to make this app more helpful and more visually appealing.

What's Next

• include sounds
• dynamically generate chord progressions (I've got an idea — maybe this will be my next hack)
• use this to make a song — gotta finish what I started last time!

Built With

• flask
• nltk
• python
• sklearn