“As an environmentally-conscious consumer, I want to dispose of my waste properly, but I often don’t know where to start – and I know I’m not alone,” Elvin said of how he and Maxim came up with the idea for their app. “For example, where I live, we can choose between several different recycling companies that have their own rules. Then, we also have municipality rules, county rules and state rules that we have to adhere to. We all want to do the right thing, but it isn’t easy to know what the right thing is.”
With the Trashly app, users can learn how to responsibly dispose of waste and be encouraged to recycle more.
Leveraging powerful machine learning capabilities built directly into Android devices, as well as the device’s camera, Trashly users can scan their items into the app. By analyzing the material composition through a custom TensorFlow Lite model, the app then makes a recommendation for how to dispose of the item based on the user’s location. For items that can’t be recycled at home, Trashly finds and recommends nearby facilities where users can drop off their items. For items that cannot be recycled, Trashly makes recommendations for how to compost the item or where to donate it, to ultimately prevent as much waste from going to landfills as possible.
The app even remembers items users have scanned previously, making it easy for users to simply search the item in their history rather than having to scan an item each time.
Elvin, Maxim and their team didn’t stop there though – they wanted to make sure that the app is as fun as it is useful. Together with EPAM’s game tech team and a game designer, they worked together to brainstorm, design and implement a game that reinforces common recycling guidance. The game element was developed using Flutter technology that was then integrated into the native Android application developed using Kotlin programming language.
“We wanted the app to have both an educational value and an entertainment value,” said Maxim. “With the integration of a simple game, we gave the app a bit of an addictive element, which will hopefully make users more excited about being environmentally-responsible consumers.”
The process of designing the app wasn’t as simple as its clean UX would have users believe. The team faced two challenges in developing the app: gathering the amount of data needed for the app to run effectively and then ensuring the app’s storage size was suitable for downloading and use on mobile Android devices.
“With each US location’s vastly different recycling rules, we needed to collect a large amount of data to make sure the recommendation engine worked properly,” Elvin said. “Once we collected that data, we started with a model that had the ability to recognize 11,000 items, leading to the app being 600 MB large – far larger than the 40 MB limit the Google’ Firebase ML Kit has for its ML models.”
Working with EPAM’s data scientist experts, they were able to condense the model’s size without compromising any of the data that makes the app so comprehensive and useful for consumers.
And, while there are other recycling apps on the market, the fact that the team developed Trashly natively to Android gives users a better experience.
“The UI/UX component of this app is what differentiates this app from others,” said Maxim. “When you develop for several platforms at once, you often have to make tradeoffs that you otherwise wouldn’t have to. By developing this app specifically for Android, we were able to develop a clean, simple interface that’s more easily accessible for all users.”
The app is currently being piloted in California, Illinois and Pennsylvania with the goal of expanding to more states and countries in the coming months. The team also has plans to develop the app for iOS users as well – without losing those essential UI/UX features.
“We’re excited to see this app come to life thanks to Google’s Android Development Challenge,” Elvin said. “We can’t wait to see how this solution makes people’s lives easier and, ultimately, makes the world a better place.”
Ready to recycle? Download Trashly here.