We made an app that takes your Spotify user data and uses machine learning to craft a playlist made up of similar songs! This project was fun, challenging, and intensive. Over the past seven weeks our team of data scientists:
-
Aggregated more than 700,000 songs to train our algorithm on.
-
Built a machine learning algorithm to predict similar songs based off of acoustical data and metadata.
-
Built out a PostgreSQL back-end hosted on AWS RDS.
-
Used that backend to store user and song data for further analysis.
-
Scraped additional data from Spotify’s API to use with our model training.
-
Migrated the entire back-end to to the suite of technologies offered on AWS, including Sagemaker, ECS and EC2, Elastic Beanstalk, and S3.
I love seeing what teams can accomplish when they push themselves!
https://www.sound-drip.com