Inspiration

We wanted to target a really large variety of customers, so a product that solves environmental issues would affect every single person on the planet. We learned about the dire issues regarding recycling in Asian countries and how much recyclable waste is ending up in landfills due to people’s ignorance. In fact, out of 75% of trash that’s recyclable, only 35% is actually recycled.

What it does

Our app will potentially allow the user to take a photo of any item they wish to dispose of. Our developed ML software will analyze the data and identify all aspects of the image. The software will then categorize the item based on its facets and will inform the user how to dispose of it or reuse it in the most effective and easiest way. Our app has multiple features that help users meet new people, do something good for the Earth, and efficiently throw their trash away in the correct bins.

How we built it

We used the Google Colab Notebook to develop an image recognition software. We collected over 1000 images to train the model. We actually got amazing results for the lack of time and experience we had. We researched immensely and tried to understand the concept the best we could. We went slowly and developed each part one by one. It was very exciting when we got something to work as we put a lot of effort and time into this project. We also wanted to simultaneously develop an app to implement this software into. Although we wanted an IOS app, it was too complex and we were too inexperienced in the area. We gave it our best shot but then settled for an android app built in Android Studios. We weren’t able to find a way to link the ML into the App but it is an obstacle we are deeply excited to overcome in the near future.

Challenges we ran into

We didn’t have a lot of experience with TensorFlow or any type of ML software. We had to do a lot of initial research to learn about how to use it effectively for our intention. As we were learning about how to use Google Colab, we made a lot of mistakes, but eventually learned from them. It was a hard and grueling trial and error process, but we persevered through the challenges and made our product as accurate as we could. In addition to the ML software, we also had to newly learn how to code an app using Android Studios. It was intimidating and frustrating at first to learn how to use, but eventually, we got it to work and built most of the pages that we had in mind. We spent a lot of time trying to figure out the code to link a camera to the app, and we were close to giving up on many occasions, but when we finally got it working, it was a very huge and exciting moment. With the help and advice of the mentors, we learned that an iOS app would be the best to work with our ML software, and we spent hours of effort into that but unfortunately, we were not able to integrate those together yet.

Accomplishments that we're proud of

We have a working ML model that identifies images at a satisfying accuracy rate. We were deeply surprised and proud of ourselves for successfully achieving a goal we didn’t think we would. We also learned so much due to the trial and error process, along with the immense amount of research. We worked very well together and divided the work up fairly and sufficiently. We had an efficient productivity rate and got so much more done than we had aimed to.

What we learned

We learned that it is important to always save your work every few minutes or so. We also learned a lot about how ML works and what it is used for. We discovered many intriguing things we can do on Android Studios to develop an aesthetically pleasing and user-friendly app. We also learned that we have to devote more time into learning how to develop an IOS app to combine the ML and app together. Another thing we learned is how much thought goes into planning and building a project such as this one. We learned that is it important to be patient and open-minded while working on such an error-prone project; things don’t always go the way you want to but you shouldn’t give up.

What's next for Upcycle Environmental

For our future steps, we want to fully develop the app. To do so, we would need to combine all the different aspects that make our app successful. We would need to implement the camera and image recognition software into the app. We plan to make an IOS app and an Android app to increase the amount of users. Another plan we have to further enhance Upcycle as an app and brand, in general, is to partner up with multiple nonprofit organizations. Primarily, we would like to partner up with organizations in our area. Obviously, in future years, we want to be partnered with nonprofits around the world. Our plan with the nonprofits is to give back to the community. Essentially, every 100 pictures taken by a user, one tree is planted somewhere in the world. This motivates our users to take as many pictures as they can and give back to make the Earth a better and healthier place. Our events page can also contribute to this. Every 5 events attended by a user or every event hosted by a user results in a tree planted somewhere on Earth. Our end goal is to have a fully functional, aesthetically pleasing, and user-friendly app, along with a successful partnership with an organization that believes in and supports our cause.

Built With

Share this project:
×

Updates