PoKoupon

Inspiration

We were inspired by the daily task of maximizing credit card rewards. With multiple cards providing different categories of cashback rewards, it is simple to forget which one to use when you get to the checkout. We aimed to make it easy to consistently select the best card. Also get coupons from local businesses!

What it does

With Pokoupon, you could discover which card offers the largest return when spending at a merchant.

Integrates with the Leaflet API to find businesses in your surrounding area
Maps merchant categories (like dining, groceries, travel) against your pocketbook
Provides an evaluation of the card that gives you the most cashback or points

How we built it

Frontend: HTML/CSS futuristic retroconsole based design
Logic: Modular JS for the UI managing wallet services, reward calculations and location functionality
APIs: Leaflet API for location and business categories
Hosting: Light server via python3 -m http.server

Challenges we ran into

One of the key issues was merging a simple demo UI for a wallet with a fullstack rewards database based on localStorage to create an integrated system. This required lots of documenting and refactoring.

On top of all that, we struggled with mapping Leaflet's place data into predefined credit card reward categories. Merchant types do not always map within specified categories (dining, travel, etc) which prompted us developing a manual mapping process and debugging troublesome edge cases.

We also faced challenges in maintaining accurate geolocation. Geolocation was often inaccurate because of browsers holding onto and caching previous coordinates or failing to provide geolocation altogether.

Using spending API keys with Leaflet also called for some devoted security / proper functionality.

Lastly, hackathon time constraints led us to dismiss some others exciting and big ideas we had, especially ideas around using ML for personalized recommendations. Ultimately, we had to stick to the basics for our presentation: location --> category --> best card (for the user). In a few cases, we had to bring attention to function, aesthetics, and style while combining our work together across each contributor. This can require some attention to flow and organization for better documentation and communication.

Accomplishments that we're proud of

Created a functional prototype with geolocation, maps, and a reward optimization engine.
Created a clean, gamelike GUI that presents finance in a non-threatening and fun way.
Developed experience tying multiple subsystems (maps, wallets, optimization) into one app and doing this pretty quickly.

What we learned

The role of transparent data flow is very important at different overlapping modules.
Experience using Leaflet API and dealing with categories for clients.
In a hackathon env, sometimes just a simple deployment is all you need (Python HTTP server).

What's next for PoKoupon

Extend to include all the major card issuers and networks, while ensuring rewards are up to date and accurate.
Support user persistence (login + synced wallet).
Begin using machine learning to find the best new cards to use based on user spending.
Release a simplified web or mobile app with a full domain (like pokoupon.app).