Taste Tracker

Ringo Starr
Eclectic Radio Worker
Batman.
Record Store Jag

Inspiration

We all are big fans of Spotify, and use it in our daily life. Everyone's favorite time of year is when the Spotify Wrapped comes around, a summary of your past year of listening. Many webapps exist to give an analysis of stats, but we decided to have a little fun with it.

What it does

The user signs into their Spotify, allowing our app to read their top 50 songs. These songs are sent to four different Hugging Face prompt that uses different criteria to judge the users music taste.

How we built it

Spotify API is language agnostic, but we used Spotipy for seamless python integration. Once the Spotify end was handled, we also created a function and prompts for the Hugging Face LLM calls. To put it all together, we used HTML/CSS and Flask to get a webapp made. We also used Aesprite to create original sprites for all of the judges.

Challenges we ran into

One big hurdle was deciding how we would make this a webapp. The first few hours of development were spent trying to figure out what techstack to use, making sure we didn't develop in a way that would bite us later. Interfacing with Spotify and Hug Chat proved to be a bit of a struggle. The returned datatypes were a bit obtuse looking from the outside in, but once we dug into a debug we were able to extract all relevant data.

Accomplishments that we're proud of

We really love the responses that our judges give. We find them to be quite humorous, and hope you will too :)

What we learned

Our big takeaway is basic web development skills, from the basic skeleton of a program, to HTML/CSS work. Much of the work was split up between team members, but all of us learned a little about what the others were working on as we went. Many of us were familiar with API usage, so the frontend is where most of our problem solving occurred.