Inspiration

Here at Digiceipt, we sought to find a simple and revolutionary way to manage receipts online. We also wanted to create a website with a straightforward design that anyone could use. With that in mind, we created MyDigiceipt.com

What it does

The app uses a series of pictures, either uploaded or take via a camera, to create budgets, charts, and deliver financial insights. This will help an individual easily manage their payments and can enable small businesses by automating their profit tracking. Overall, by combining many new technologies such as computer vision, we will help make the difficult and often impossible task of managing all your receipts an easy process.

How we built it

The stack we picked was full stack JavaScript. This includes the popular JavaScript tools like Node, MongoDB, and Express. We also used front end tools like bootstrap to design a responsive website. Because it is responsive, it serves both as an app and as a website. We also used Microsoft Azure to easily handle the text processing.

Challenges we ran into

The actual process of scanning a receipt and getting the appropriate data was really difficult. Thankfully, using Azure we were able to alleviate much of the initial difficulties we had. We also had trouble creating a unique way to sign.

Accomplishments that we're proud of

While working on this project, we were proud to have implemented Microsoft’s Computer Vision API to scan the receipts uploaded. Coming up with a creative way to authenticate a user without using a traditional password. As a team with a majority of first-time-hackers, we were able to have a platform to advance our web development skills outside of the classroom.

What we learned

We learned how to work with Microsoft’s Computer Vision API. We learned how to implement photo taking capabilities into a website/app. We learned how to work together in a short amount of time to create an amazing product!

What's next for MyDigiceipt

Although it was not possible to implement a machine learning system that could parse any receipt we handed to it, we look forward to working on training a neural network to achieve this goal via crowd sourced submissions through our own website. In addition to this, we would like to implement a feature that cross-references a user’s most commonly purchased items with coupons. We hope to also make the coupon system compatible with Google Maps Platform to provide users an easy way to locate where they can apply their potential savings.

Share this project:
×

Updates