Waste has become one of the World’s biggest problems - The amount of waste generated has tripled since 1960. As a Zero-waste oriented founder, our team member Max tried to do things right at his company. They have different garbage bins, like plastic, paper, glass, organic, other. And Max thought they were doing everything right. But after inviting a Zero Waste expert for a Workshop - Max was shocked. They were doing so many things wrong. And there are so many types of garbage! It turned out to be a big problem for other communities.

In 2017 China, as part of an anti-pollution crackdown, even announced it would stop importing most used plastic and paper because it’s not properly sorted.

That’s why we've created Oscar - a Zero Waste Expert for your community

What it does

Oscar solves 2 problems:

  • Consumers don’t know what can be recycled and how
  • Consumers are not motivated to recycle and reduce waste

How it works:

  • Oscar is using the power of communities to bring motivation to reduce waste by gamification (your community and you have a rank and as you progress, you move to the top)
  • Oscar gives you daily tips on how to become “Zero Waste” (recycle, reduce) based on your stage;
  • Send a picture when purchasing and Oscar will tell you if the product is recyclable or waste.
  • Send a picture when throwing something out and he will tell how to recycle or if it’s waste?
  • Oscar tracks what you throw out vs what you recycle
  • Community Managers can create rewards for the community members that reduce the most waste and increase their recycling. (Recycling Bank, Private Recycling companies, City Government, etc.)

How we built it

  • We built the application using java and integrated with Facebook Messenger API to create a chatbot that would interact with the Community Initiator (Manager) and Community Members
  • We built a way to handle image uploads including multiple filetypes using node and javascript
  • We built an improved algorithm and image recognition model on top of the Google Cloud Vision API using node and javascript as a backend micro-service the chatbot can chat to.

Challenges we ran into

  • Preparing the experiment of running our chatbot in real-life here at hackathon
  • Training the algorithm to determine different types of trash - this was actually quite challenging as handling multiple angles for the same item - definitely an area we want to improve going forward
  • During the launch on the hackathon, we ran into the problem, that people were sending not what we expected: pictures of their friends, teammates, some random pictures not related to waste.

Accomplishments that we're proud of

  • We ran an experiment, which helped us to find and fix the weak spots
  • We attracted 50 users (2.35AM, April 30)
  • We were able to implement the solution at the hackathon
  • We also changed our behaviors of how we're consuming at the hackathon based on our own app and from what we learned about waste

What we learned

  • We learned that to build an accurate object detection and classification via open APIs is a challenging and time-consuming task
  • We learned much more about waste problems and how we can optimize our waste in everyday life

What's next for Oscar

  • Increased personalization based on the answers you provide Oscar (personalized advice based on how you answer, progress and interact with Oscar)
  • Leaderboard → improved gamification features to help motivate community members to reduce their waste and recycle more
  • Goal setting for individuals → helps with motivation
  • Joint accountability with other members of your community - each sets goals and encourage each other to succeed together
  • Location specific information about how and what can be recycled → what types of recyclables are accepted in your area and where to bring them to recycling them
  • Petition government for improving recycling programs (when an item can’t be recycled)
  • Open Data: Share consumption and waste reduction with policy makers

We also think there are some potential business models to sustain the development and scaling of Oscar to the world:

  • Sponsorship for recommended alternatives that are ecologically friendly and help you reduce waste, for example:
  • Image Trained Data set for Commercial Waste Sorting at Materials Recovery Facilities

first link - final product; second link - experiment chatbot ran during the hackathon

Built With

Share this project: