img

We put 4 developers in a room for 48 hours. You won't believe what happened next...

2020-04-28

Escape rooms

During the "Nordea Innovation Challenge," we aimed to reinvent banking by creating an app using Python/Django for the back-end and Angular with Ionic for the front-end. Our app was designed to scan receipts, link them to products, and offer savings recommendations. By integrating the Nordea API, we accessed transaction histories and built a back-end for image handling and data anonymization, along with an OCR service for receipt parsing.

48-Hour Innovation Hackathon: What Our Team Achieved

One idea, 48 hours, GO

So, we participated in an Innovation Hackathon called "Nordea Innovation Challenge". The challenge was to "reinvent banking in the digital age". Quite simple, right?

Our strategy was simple:

  • We wanted to build an app

  • We wanted to use our favorite stack (Python/Django for the Back-End, and Angular for Front-End).

The only thing we needed was a good Hybrid-App Platform. After browsing the web for a while we chose Ionic - "The beautiful, open source front-end SDK for developing hybrid mobile apps with HTML5."

Some brainstorming later we found some problems that we would like our App to solve:

  • Is it possible to take a photo of my receipt and automagically get a list of products that I have bought?

  • Can I open an app and search for products and easily find the receipt connected to that specific product?

  • Can I get recommendation on how I could save money by NOT buying milk from, say, 7-11 and go to my local super market instead?

By using the Nordea API we had access to users transaction history. We could also fetch some other resources, but those weren't applicable to our user stories.

So, what were the core components?

  • A Mobile App

  • Integration with Nordea API

  • Build a Back-End to handle pictures and anonymize data about user transactions

  • Build or integrate a service to handle the raw optical character recognition (OCR)

  • Build a service that analyzes the result from the OCR-service and parses products and prices from the receipts.

Easy enough? Here's a simple architectual overview of how the end product was built.

Illustrated architectural overviewIf you want to know more, feel free to contact us through our website: www.willandskill.se