Berry Disco

Community player in action, songs play in a loop so vibe never dies
AI recommends songs for specific theme

Inspiration

Our journey began with our past experiences at hackathons and insights gathered from previous organizers. We recognized a common challenge: the hassle of manually managing song requests during events, which often disrupted the flow of music.

What it Does

Disco Berry is a web app designed to streamline the music experience at community events. Organizers can create an event using a unique code, allowing users to access the event page and submit their song requests. If the playlist runs low or is yet to be filled, our smart shuffle feature kicks in, playing music tailored to the event's theme. This ensures a continuous vibe from start to finish.

How We Built It

We developed the web app using the Streamlit framework for an intuitive frontend while leveraging Python for the backend. Our integration of OpenAI's embeddings and chat completions, along with structured JSON outputs, enabled us to harness the power of a language model effectively. For our database, we utilized Azure Cosmos DB, which served as the backbone for storing events, event codes, and user song inputs, providing robust data management.

Challenges We Ran Into

One of the primary challenges was the frontend design. Initially, we hoped to find a team member with React skills to enhance our app's aesthetics, but we ultimately relied on the Streamlit experience of one of our members. The most significant hurdle was initially setting up Azure SQL Database; however, after realizing the complexities and authentication issues we faced, we pivoted to using Azure Cosmos DB, which ultimately worked better for our needs. Additionally, we attempted to utilize** Azure Web** App Services for deployment but ran out of time to implement it fully. Thankfully, we overcame most of these challenges through persistence and teamwork.

Accomplishments We're Proud Of

We take pride in the efficacy of our LLM procedures, which provided accurate song recommendations on demand. Additionally, the seamless integration of our app components and the creation of a user-friendly experience have been major achievements.

What We Learned

Through this project, we gained invaluable experience in cloud database management, frontend development using Streamlit, and leveraging AI for enhanced user interaction. We also honed our problem-solving skills while navigating various technical challenges. Additionally, we gained deeper insights into Azure Cloud Services, further enhancing our understanding of cloud-based solutions

What's Next for Berry Disco

Looking ahead, we aim to improve the frontend aesthetics and functionality. Potential features include a voting system to rate song requests and a counter to track the most popular songs. We also envision expanding the app’s capabilities to accommodate different event themes and enhance user engagement.