SmartCart

Inspiration

When a shopper enters a store, a common problem they face is the paradox of choice: given so many choices, it becomes increasingly difficult for a shopper to settle on one. The goal of our application is to solve this problem.

What it does

After turning on the mobile application, the user can scan a food-related product. Afterwards, the user receives nutrition information regarding that product, such as the pros and cons of the nutrients inside the item, and ratings from the user base. The user can then compare that with a similar item to decide which item to buy. When the user is done, he or she can leave a review to share with other shoppers--and all of this can be done with just a few taps.

How we built it

Our application is front-ended with React Native, a framework similar to React. React Native essentially transpiles Javascript into binaries that are usable through all platforms. Our back-end is built with Firebase's Cloud Firestore, an innovative database system built by Google. We used an open source machine learning API written in python to build image recognition for nutritional information; with just a scan, all the information on a nutritional label on a product is sent to our code. We host the API running this program on a virtual machine hosted by Google Cloud, and we send requests to the virtual machine when we need to process an image.

Challenges we ran into

A major milestone we had to pass was that we were not that familiar with React Native. Although we had worked with similar frameworks in the past, working with React Native was still virgin territory to us. Another important problem we had to address was hosting the machine learning framework over Google Cloud suite. There were many technical difficulties we had to deal with this system, so it was quite a hurdle for us.

Accomplishments that we're proud of

One accomplishment that we were proud of was creating a working nutrition label recognition system. Setting up the network between our virtual machine and our mobile application was an arduous process.

What we learned

Through this Hackathon, the one thing we learned the most probably concerned React Native. Going in, all of us had little to none experience with the framework, so we were able to explore this topic greatly during this Hackathon.

What's next for SmartCart

A long-term focus of this project would be monetization. We can see two different approaches to this; a) through ads, or b) through supermarket sponsors; if supermarkets push their deals and discounts to our application, they gain more marketing and we are able to gain more users. Although these ambitions are quite a long ways away from the working model that we have developed at this Hackathon, we believe with time, such a hope can be realized.