Inspiration

As busy yet broke college students, we needed an efficient way to track our spendings.

What it does

Takes an image of a receipt, obtains the spendings for each purchase category, and documents all of it.

How I built it

We first created a GUI for the user to photograph the receipt. Using the Google Cloud Vision API, we transformed the image into text, which can then be processed by Python into data. The final output is returned back to the GUI.

Challenges I ran into

  1. Tried to use TensorFlow at first, did not succeed.
  2. The Google Cloud Vision API returned data in a non-ideal format. We had to come up with further programs to turn that into useful data.

Accomplishments that I'm proud of

Completed everything in 12 hours on 3 hours of sleep.

What I learned

Machine learning is hard. Thanks Google!

What's next for sorted(receipt)

Nicer GUIs, ability to add store locations and breakdown of individual items bought.

Built With

Share this project:

Updates