Spending Saver

Inspiration

People tend to buy things before they think and even smaller purchases all stack up. So we are bringing awareness to the people to help put into perspective how much they can save if they don't impulse buy.

What it does

The app takes in the users input of items they have bought and will classify them based on essential and non-essential. Then our logic engine will calculate how much they can save in the future if they cut back and don't buy the non essential items. Then the predictions and data will be shown to the user to let them know that spending $6 on Starbucks daily adds up quickly and may even cost them a new high end computer a year.

How we built it

We built a modular Python backend focused on flexibility and scalability, handling core financial logic like normalization, projections, and trends while returning clean, API-ready responses. The system uses Snowflake Cortex AI for prompt-based spending classification with structured JSON outputs, supported by a data normalization layer that converts different time frequencies into comparable daily values. A caching system reduces repeated AI calls for better performance and cost efficiency, and the architecture is fully pluggable, allowing easy switching between providers like Gemini and Snowflake without changing core logic.

Challenges we ran into

We ran into many challenges some of which are figuring out how to use swift, getting snowflake cortex to work, and using snowflake with the backend. These problems all occurred because this was the first time any of us have used these frameworks/languages or programs.

Accomplishments that we're proud of

We are pretty proud of making a nice looking swift interface, getting the database setup with snowflake, and getting snowflake cortex working. All of the problems we faced and conquered is what we are proud to have accomplished.