QuickSort

Snap, Sort, and Save the Planet! 🌍 With just a click, make your trash talk eco-friendly with our app. Don't waste another moment—join the clean scene team today! 📸♻️

Comment

Inspiration

  • Manual sorting of garbage is a difficult and expensive process and the global waste for recyclable materials has significantly increased over the years.
  • It is projected by the World Bank that in 2050, the world will generate 3.4 billion tons of waste per year which is a 70% increase compared to 2023.
  • Our strong desire to reduce the negative impact of waste and maintain a sustainable environment in the next 100 years brought us together, brainstorming methods for automated garbage sorting to improve the recycling process.
  • We shared knowledge about machine learning, deep learning, Google Cloud, and full-stack development skills and we decided to design a mobile app that can automatically quick-sort the garbage with the support of machine learning models and lead the user to the nearest recycling places on Google Maps.

What it does

  • Users can scan everyday objects with the scanner in QuickSort and will immediately sort things into various categories with the support of the Machine Learning Model.
  • The Maps feature will lead the users to the nearest garbage recycling places on Google Maps to improve the recycling process.
  • With the support of Google Cloud, QuickSort will record the total amount of garbage recycled by the users in the past year/month/week and display the data on line charts and pie charts.

How we built it:

  • Machine Learning: PyTorch, Deep Learning
  • Hosting: Google Cloud, CockroachDB
  • Frontend: React Native
  • Backend: Node.js, Express.js, TypeScript, PostgreSQL
  • UI Design: Figma

Challenges we ran into

  • Flask server setup to connect with MongoDB and Google Cloud SQL.
  • Implementation of authentication with Google.
  • Speed of classification using our models.

Accomplishments that we're proud of

  • Trained numerous models using PyTorch with over 20k+ images.
  • Set up the Node.js server to connect with CockroachDB hosted on Google Cloud.

What we learned

  • Mobile development using React Native for device cross-compatibility.
  • Learned the process of connecting to various database services (Google Cloud SQL, MongoDB, CockroachDB)
  • The process of setting up a server in Flask and Express.js.

What's next for QuickSort

  • Live-feed detection for rapid detection speed to improve user experience.
  • Global recycling management system to track total garbage recycled on a global scale for further research.
  • Implementation of a bonus system based on the amount of garbage a user recycles to encourage them to reduce waste production everyday.
Share this project:

Updates