Jam Out

With coding designed to match musicians to other musicians with similar musical tastes, Jam Out is the perfect way to connect! Just click on our sign in page and fill out your information to join.

Comment

Inspiration

Since the pandemic harmed the live music industry, we wanted to create a network where Texas A&M musicians can meet and "jam out!"

What it does

Jam Out recommends similar musicians to each other based on their music taste, using a machine learning algorithm and advanced web design.

How we built it

First, we began by implementing a tensorflow deep learning algorithm with embedded systems. We created the sample users using the Spotify API for possible songs and genres as well as a rating algorithm based on their chosen genres. Then, we trained the model based on the user data, genre data, and rating scores. Ultimately, the algorithm returns a group of the best matching users. In order to display the results, we consolidated the user profiles and recommendations into a website using HTML, CSS, and JavaScript. We linked user pages and subsections with clever file organization, and we utilized a base template with navigation and background settings to create a fluid, consistent format that is all-encompassing yet user-friendly.

Challenges we ran into

  • Testing and tweaking the algorithm (analogous)
  • Understanding the Spotify API
  • Formatting and creating CSS
  • Consistent navigation bar
  • Working and modifying bootstrap templates

Accomplishments that we're proud of

  • Making a consistent navigation bar
  • Installing tensorflow and the appropriate environment
  • Scraping data from the convoluted Spotify API
  • Adapting the tensorflow algorithm to an analogous situation
  • Connecting the webpages together seamlessly

What we learned

  • HTML tag system (buttons, navigation bar, search bar, linking, images)
  • CSS styling and overwriting system
  • JavaScript functionality and csv reading
  • Using the Spotify python API, Spotipy
  • Learning how the ML works (batch size, epochs, learning rate, KNeighbors, embedding systems)
  • Shifting the algorithm to the situation (data setup, variable modification)

What's next for Jam Out

  • Hosting the website onto a web service like AWS
  • Storing the data on SQL database and accessing through Promise requests
  • Adding custom JavaScript animations and dynamic capabilities
  • Further configuring the model to be more accurate
  • Using real user data
  • Changing the font size for the visually impaired
  • Optimizing webpage linking
  • Adding venue and messaging functionality

Built With

Share this project:

Updates