Safe Packaging

• 

  Our Model's predicted packaging material for one scenario. Each time the dimension and Packaging material changes to optimize.

• 

  Our Model's predicted packaging material for another scenario. Each time the dimension and Packaging material changes to optimize.

• 

  We created the data ourselves and had to spend a chunk of our time doing so, we generated 1000 lines of data.

• 

  The recent returns of the customers helps the ML learn further and uses the feedback to do so.

• 

  our about us page

• 

  home page

• 

  front end code snippet 1

• 

  front end code snippet 2

• 

  Our data collection for the item

Inspiration-

Our inspiration was a package one of our members received a few months ago. The package was for a mouse from Amazon. The packaging used was a huge cardboard box but only 10% of the box was being used and there were also a bunch of packing peanuts and bubble wrap. We had to throw all the packaging away except the cardboard box because packing peanuts and bubble wrap aren't recyclable. We then realized that packaging for shipments in the country is really bad for the environment and inefficient, the sheer numbers are crazy, in fact 30% of all waste in the US is just from packaging, and of this 91% of it ends up in a landfill or the ocean. By making the packaging optimized we would save a lot of trash from harming the environment and we would improve the environment and reduce trash pollution

What it does-

Our website uses a machine learning model to optimize packaging for shipments based on fragility, dimensions, weight, and the users' budget. The goal of our website is to drastically reduce waste in packaging by giving users more environmentally safe packaging methods without making them spend large amounts of money.

How we built it-

We built our website using HTML, CSS, Javascript, and Python. Two of us focused on the front end and UI of the website. We used HTML, CSS, and Javascript for this. One of our teammates focused more on our machine learning model, he used Python to build the machine learning model.

Challenges we ran into-

The challenges we ran into were making the AI. We had to use multiple complicated algorithms and functions to accomplish this. We didn't have a dataset so we had to make one using these mentioned algorithms and functions. This took the majority of our time because it is deep learning, this takes a lot of data to be accurate. It was worth it however because our accuracy reaped the benefits. We also had to make an API for the ML model because we needed to connect the model to our front-end.

Accomplishments that we're proud of-

We are proud of making a successful and accurate machine learning model. It was our first time doing this and we dove in headfirst without ever trying it. We are very proud of ourselves for making it work on our website.

What we learned-

We learned how to make a machine learning model to successfully solve a big problem. We also learned to connect a UI with a backend machine learning model and to present a user with options of what to do based on our model.

What's next for Safe Packaging-

The next step for Safe Packaging is to build our product for other countries using their systems of currency and measurements.

Built With

• css
• deep-learning
• flask
• html
• javascript
• ml
• python
• tensorflow