staryeast

StarYeast
Starry Yeast
Yeasty Stars

Inspiration

What if the music we listened to adapted to our environment? What kinds of music would you listen to when its raining, sunny, or freezing?

What it does

Based on live weather conditions, we generate playlists to listen to.

How we built it

Using Google Trends, we downloaded data on the popularity of songs over time. We then matched each date to weather data to see how song popularity changes depending on weather conditions. Using an Kaggle dataset, we matched songs to their characteristics including danceability, tempo, explicitness, speachiness, and more. We trained a LSTM model to predict song characteristics based on the weather. We then used linear regression to match the predicted song characteristics to five songs out of a database of around 1 million.

Challenges we ran into

Switching projects a million times: worked on procedural music generation, AI music generation, and more. Decided Saturday midday on this project.
Getting daily data from Google Trends
Making machine learning model do something
Data wrangling
Library installations
Klaus lights don't work plz fix.

Accomplishments that we're proud of

Clean interface using React.js for the first time, used machine learning to make predictions for the first time. Got both the front and back ends working in a short period of time.