Inspiration
Music recommendation systems have become an essential part of our daily lives, yet current systems often fall short at providing truly personalized and dynamic recommendations. These systems frequently struggle to find the right music for individual users and tend to be static in their recommendations.
Recognizing these limitations, our team was inspired to develop a unique approach to music recommendation. Our goal is to create a system that doesn't constrain users' tastes while still allowing them to explore new musical horizons. This system is carefully curated to individual preferences, ensuring true fine-tuning of recommendations.
What it does
The app determines the user's mood based on their conversation with its chatbot. Then, it curates a recommendation list of songs based on the given mood where users can use interactive features like live play/pause controls and mood-based "likes," so users can engage with their favorite tracks. Additionally, integrated chatbot support enhances the experience, making music discovery smarter and more intuitive than ever.
How we built it
Our application's front end was built with React and our backend with Flask. Our backend utilizes data analysis and machine learning to determine optimal song recommendations. We also made our chatbot with Bedrock's Titan model. The model performs sentimental analysis on the user input. We hosted it with AWS because it reduces communication friction between backend and the chatbot. We built machine-learning model along with data cleaning pre-processing and analysis to get our recommendations.
Challenges we ran into were logical challenges with data analysis along with machine learning recommendations with software development which includes Spotify API and database management.
A challenge we ran into was using Spotify API to scrap the playlist where the API was not authorizing it.
Accomplishments that we're proud of
Proud of creating a fully working backend and frontend application.
What we learned
We learned how to deploy an application with AWS, how to use React to develop a high-quality front end UI, and how to combine multiple machine-learning models
What's next for Recommendation App
Fine-tuning certain model hyperparameters, increasing integration with Spotify, deploying back-end with dynamically changing AWS resources.
Log in or sign up for Devpost to join the conversation.