We were inspired by the recent explosion of #TrashTag, which featured people from across the world cleaning up the environment and posting it on social media. In this day and age, where social media has such a large reach, we thought that this application would be a great way for people to be able to show others the good they have done for our environment.

What it does and how we built it

We built a multi-technology iOS app that uses many different frameworks. We built the iOS frontend with Swift which uses Google Cloud Vision API in order to classify camera input from the iPhone and determine if the item scanned should be recycled, composted or trashed. We used python to scrape a few web pages we deemed helpful in building our set of values for recyclables, compostables, and trash. We then used a Firebase Database to connect users to their personal recycling/composting/trashing statistics. We also used Firebase Functions in the backend in order to pull counts from the Real-time Database and generate Snapchat Customized Stickers with that users stats. The iOS app now allows the user to prepopulate their next snap on Snapchat with the Snapchat CreativeKit. Primarily languages/technologies used were Swift for iOS app design, Javascript and Python for the backend, and Sketch for App and Sticker Design.

Challenges we ran into

Snapkit integration Swift and web language integration UI/UX of application Generating a Snapchat sticker template Learning how to deal with parallel programming issues Fatigue

Accomplishments that we're proud of

We are proud of the fact that despite a shorter time frame than we are used to, we made a full complete app and learned how to integrate Snapchat into our application, which is something none of us have ever worked with before. On a related note, we are also proud of the integration of all the different technologies we used. It was pretty tough, and we did run into some bugs that were pretty demoralizing, but ultimately we were still able to make something we’re all proud of.

What's next for ScrapSnap

We want to improve the way we classify the items. Right now, it’s definitely a very rough classifier because of the short time span. However, if we were able to maybe implement machine learning through picture learning, the classifier would be a lot better, which is definitely something that can be done in the future.

We also were thinking about implementing a global leaderboard system amongst all users to make it a friendly competition to compete with people around the world.

Built With

Share this project: