HamsterWallet

An AI assisted budgeting app for students

Comment

Inspiration

Our randomly assigned team name gave us the basis for our app - allowing us to brand ourselves around the Hamster. We chose to create this project as it's very relevant and useful for people like ourselves who need the extra assistance managing finances whilst studying at university. This meant we could create features which we knew would be useful for ourselves.

What it does

This app, designed for mobile use (or responsive mode on desktop), is used to assist students with budgeting. The app relays to the student their weekly expenses based off spendings they've inputted. These transactions are auto-categorised by a Large Language Model (GPT-4) and displayed in the donut chart on the main page. This way a student can easily track weekly expenses and budget the remaining money they have.

The app also features a login and registration system which both communicate with a database to allow the user to login in and register. (An appearance is also made by "Henry B Hamster" on the home page).

How we built it

Frontend: Vanilla JavaScript Backend: Django, SQL

Challenges we ran into

There were many design challenges we faced on the frontend. Another was choosing the model we wanted to use in order to categorise the various transactions taking place effectively. We struggled with team scheduling conflicts but over the 24 hours made the app completely from scratch.

Accomplishments that we're proud of

  1. Our logo (Henry B Hamster).
  2. Completing the project.
  3. Working with new people and new technologies.

What we learned

Lots of Django, git, and how to cause as many conflicts as possible...

Demo Account

Email: demo@durhack.com Password: demo

What's next for Hamster Wallet

If we were to continue with this, we would like the user to link their bank transactions and automatically categorise these transactions so that the user wouldn't have to enter these manually. We would also like to look at putting more customisable constraints on budgeting each category of spending specifically. We could also use an LLM to recommend these values based on the user's needs.

Share this project:

Updates