-
-
Initial view with autofill
-
Overview of analysis and some negative scenarios
-
Positive scenarios and final AI analysis
Inspiration
One of our friends has a history of very poor financial decisions when it comes to investing. The whole idea we had was what if there was some tool he had that would give him insight about why certain investments might not be a good idea, or why certain investments might be a good idea. The idea came together fully when we were presented with tracks and challenges. We saw the Gemini challenge and thought of using Gemini to have an analysis of portfolios in order to make scenarios of what might happen to your stocks in the future.
What it does
A web-based Portfolio Stress Testing application that analyzes how your stock portfolio performs under both bull and bear market scenarios. Built with AI-powered insights using Google Gemini, this tool helps investors understand their portfolio's risk-reward profile through dynamic, personalized stress tests.
How we built it
Our workflow started with making a prompt to generate code. We would use Claude AI to give us css, js, and py files in order for the app to actually run. After this, we would perform testing to make sure inputs were valid, and that the proper responses were obtained. If they weren't, we would spend time looking over the code and adding in fixes. After that, we would then think of new features/fixes, and finally prompt once again and repeat the workflow. This ended with us changing the css styling to better fit the theme before we submitted.
Challenges we ran into
None of us know Python or Javascript, and so we had to rely on Claude to generate the baselines for us. While we could look through the code and get an understanding of how something works, we would not have been able to make it from scratch by ourselves. Essentially, we could look over the code for errors, but we could not make features ourselves.
Accomplishments that we're proud of
This is our first hackathon, and so we were all very proud that we got a functional submission. We did not have high expectations going into it, but we are all very happy that we ended up getting an actual submission that could be presented to other people.
What we learned
While we may not Python, we were able to understand how the Gemini API interacts with it. Not only that, but we also got a good understanding of the yfinance library. These are the main two things we explicitly learned about while working on it.
What's next for Portfolio Performance Analyzer
One thing we wanted to add but didn't have the time for was getting a stock ticker by searching the company name. This would be a really good feature for our intended users of regular individuals, since we shouldn't expect them to inherently know ticker names, but they should know company names.
Log in or sign up for Devpost to join the conversation.