Americans represent 5% of the world’s population but generate 30% of the world’s garbage. Moreover, about 80% of what Americans throw away is recyclable, yet our recycling rate is only 28%. Especially, plastics that could be recyclable take thousands of years to decompose which might affect the sea animals like turtles as they try to eat them. Considering all the potential hazards and leaning more towards a more eco-friendly life has inspired us to make a web application that would eliminate the misuse of plastics.
What it does
The web application tells whether a certain type of plastic is recyclable or not by reflecting different wavelengths of light.
How I built it
Challenges I ran into
The first challenge we faced was collecting the datasets. Because we decided to create our own data set, we had to obtain 70 different datasets. The second challenge we faced was recalibrating the sensor. Due to time constraints, it was difficult to obtain all the databases we have collected. The third challenge we faced was working with technologies like flask, bootstrap, and firebase because it was not familiar with most of the members in the group. Therefore, we had to learn it before executing the project.
Accomplishments that I'm proud of
We are proud of making our own dataset. Instead of depending on the public dataset, we decided to make our own dataset by scanning different types of plastic materials. Moreover, we got the machine learning model to work which is used to classify the various plastics. Finally, we are proud of learning new technologies and implementing a good finished frontend result.
What I learned
The new technologies that we have learned are flask, firebase, scikit learn, and bootstrap.
The public dataset that we created and used: https://www.kaggle.com/bwolfram/spectual-data-for-recyclable-items