Quizzr

Have you ever been stuck trying to find test questions, or are looking for a fun way to play trivia with family and friends? Quizzr transforms any open dataset into a hard test or fun trivia night

Comment

Inspiration

Games such as Kahoot and Jackbox inspired us to create Quizzr. We wanted to put a twist on the genre which removed the burden of creating your own quizzes from scratch.

What it does

Quizzr is a full stack fully deployed web application which allows users to create quizzes from publicly open datasets through Kaggle, once a dataset is chosen users can host lobbies where friends can join and compete in an intense and educational trivia game. Depending on the type of question given, an informative graph will be displayed to users to help them understand the answer.

How we built it

Quizzr was built using a Python backend and React frontend built with Vite. We are hosting our backend through Railway and the front end through Vercel. We make use of several services such as Kaggle API to obtain open data and OpenAI API to generate quiz questions.

Challenges we ran into

The most difficult challenge we ran into along the way was generating meaningful visuals and questions. We opted to use OpenAI’s 4o to both generate questions and meaningful visuals, fine-tuning the prompt to obtain these was a difficult challenge.

Within the frontend, it was difficult to handle edge cases and work with WebSockets

Accomplishments that we're proud of

We are proud that within the time frame, we could deliver all the features we sought to complete, and we look forward to making future improvements to Quizzr!

What we learned

We exposed ourselves to new technologies such as WebSockets and worked with new APIs we had never used before.

What's next for Quizzr

We would love to continue developing Quizzr and fully fleshing out the project
Possible areas of improvement:

  • Refining the generated supporting visuals
  • Final tweaks and polishing to the UI

How we used AI

We all use GitHub copilot during development and OpenAI's models for debugging and scaffolding. To add, our backend uses OpenAI's API specifically 4o for its cost-to-performance ratio to generate the quiz questions and visuals.

Built With

Share this project:

Updates