BinChicken - Making Australia Clean Again

Team Name

I adopted a puppy today but I'm still here hacking

Team Members

Project Description

The problem:

Over time, Australia has seen waste generation increase, but recycling stagnate. This has significant impacts on the environment - and cost of service. The cost per tonne for kerbside waste handling is twice as high for garbage as it is for recyclables.We have also found in our daily lives that well-meaning people are confused by whether or not their rubbish is recyclable. People recycling a non-recyclable items, or vice versa, cost councils thousands.
If we can enable AND encourage more people to recycle a larger percentage of their household waste, we can save both money and the environment.

Solution:

That's why we've created BinChicken. Using machine learning, the mobile app informs users of how to recycle their items when they are unsure (see the video for a demonstration in action). It looks to track users' contributions by using data from Sustainability Victoria to calculate the impact from the items they recycle. This is displayed in relatable terms (kilometres in a car, hours watching TV, etc) along with the actual energy and emissions savings. The impact is scored and combined this into a leaderboard. Some healthy competition between neighbouring suburbs can really fuel the community in its journey toward sustainability!
In addition we have features for the user to find their closest waste management facility for additional help in recycling items that are not handled through the regular kerbside waste service.

Vision:

We hope to see a future where as much waste as possible is recycled, saving both money and the environment. We hope to see Australia lead the world into a sustainable future.

Data Story

Data To Identify the Problem

We used the Local Government Annual Waste Service Report to educate ourselves about the state of affairs when it comes to waste management and recycling. It was clear that while total waste increases, the amount recycled doesn't necessarily keep up. This told us there was a problem we wanted to solve.

Machine Learning Data

We found an image dataset that we used to start training our model. It lets us identify the material of an object (to a pretty good success rate) and we can then direct the user as to how to recycle it. In the future, we look to collect more data by letting users submit additional material and use it to further train the model.

Waste Impact

We used the LCA kerbside recycling calculator to get information about the impact of different items. Eg if you recycle 3 soda cans a week, how many kilometres of driving is the CO2 emission savings equivalent to? This lets us present the user with metrics of their impact that they can relate to more than just the kilograms of CO2.

Nearby Waste Facilities

Using the National Waste Management Database we also provide the user with information about nearby facilities where they can take items that are not handled by the kerbside waste service.