Personal Financial Advisor

Often, all it takes is an injury, a loved one passing, or something that is brewing in your finances that you don't know about that can hurt you significantly financially. The PFA is here to help.

Comment

Recent Aspirations

Recently, I got really caught up in my personal finances and setting myself up for my future. As I have gone on this journey, I learned just how bad the gap between the wealthy and the poor in regards to financial literacy has become. As I was getting ready to join this hackathon, I didn't have a clue what I was going to build or how it would be useful. Then, as I was brainstorming, I remembered that I needed to tweak something in my budget. As soon as that happened, I realized that I knew what I was going to make for this hackathon.

What it does

The Personal Financial Advisor (PFA) takes the most basic information about you (things that anybody regardless of their level of financial literacy can address) and runs it through a large language model called Google Gemini. Google Gemini takes this information and compiles a beginner friendly financial plan that is then sent back to the user for them to view and study.

How we built it

I built the PFA by first testing to see if Google Gemini was capable of delivering on this task. I was pleasantly surprised when on the first try it gave me more than what I asked for. This changed what I was willing to demand from Gemini and thus it also changed the inputs and outputs of the design model I had at the time. After testing, I began with simple setup and testing just to make sure things were working on the fundamental level (Your "Hello Worlds" if you will). Then I got to work on the backend, designing the output structure of what the model should produce. After configuring the server, I got straight to work on the front-end. I was confused by the direction I wanted to go with the front-end until I used the Figma AI designer to make a mockup that fit the aesthetic of what a Financial website should look like. Once I implemented my own version of the mockup, I then connected the server to the client, and started testing the project from there.

Challenges we ran into

Google Gemini is great but inconsistent. Yes, it gives us great explanations, data, and beginner friendly language, but it fails on even the same prompts at times. Also, it is very slow. However, for the sake of this project, I decided not to pay for the higher level plans, so maybe those would have been faster.

Accomplishments that we're proud of

I have never worked with a large language model API before, and seeing how well this project turned out, I feel as though that in and of itself is an accomplishment. Most of the other things in the project are things that I knew how to do or had an abundantly clear plan of action.

What we learned

With the right prompting and proper data, large language models like google gemini can make incredibly concise yet insightful overviews of very complex topics.

What's next for Personal Financial Advisor

Well, the PFA is not an actual advisor, so it would be unethical to ship this to market. But I think it stands as a good proof of concept for things that we all know are coming.

Built With

Share this project:

Updates