thrifti | Devpost

Inspiration

Our inspiration was our own personal experience of being a struggling college student without an affordable financial mentor, therefore we developed our own web application to be able to manage our financial budget without the need for a third-party mentor.

What it does

Our web application takes your bank statements, the average salary for the month, and your estimated amount of rent. Given this data we will compute the best budget options and utilize openai to derive meaningful insights that help the user optimize their budget. We will also organize your budget and spending information in neat and concise diagrams.

How we built it

We built it using next.js and openAIAPI by importing a sample bank statement given by one of our team members and then we proceeded to organize the data using the API. Finally, we provided various budgeting options based on the computed budget such as recipes based on your budget.

Challenges we ran into

One of the challenges we faced was properly interfacing with the ChatGPT API to convert the bank statements into data we can use. Another issue we faced was the implementation of these coding scripts using various languages some of us weren't familiar with. Furthermore 3 of us haven't had real hackathon experience so we had to learn on the fly how to operate under these circumstances.

Accomplishments that we're proud of

Developing eye-appealing bar and pie graphs for budgeting data. It was very satisfying to see the various colors given. Furthermore drawing upon one of the best LLM's availible in industry to create a comprehensive analysis engine felt increadibly empowering and shows us how much we can do with effort.

What we learned

What we learned is how to set up and configure a next.js project. We also figured out how to interface with a chatGPT API.

What's next for Thrifti

What is next for Thrifti is to finish the project by implementing a further breakdown on how to improve per category EG we had initially also planned to do a stock analysis and pantry management tool. Furthermore we also want ot expand to a mobile and finally we want to use Plaid to directly get information from the bank instead of asking the user for a csv.