Inspiration
While browsing some public data sets, we saw a well formatted and complete data set for tarot cards. After hearing from some of the trading firms at the hackathon, we thought there was a way we could combine the two topics together. Quant trading and tarot reading are at quite different ends of the world, and since quant firms are always looking for the next thing to give them an edge, we thought this was the most novel way.
What it does
Tarot trader lets you shuffle through a deck of tarot cards, picking three at random. These three cards represent the past, present, and future of the chosen asset. Using recent news, and analysis from the cards, it will pick an asset.
The asset picked will then have a decision made on it, short or long. You can choose to take the advice of the cards, or decline.
There is also a portfolio view, to see your open positions and performance
How we built it
We use a react frontend and a fastapi backend. Our data is stored in a postgresql database hosted on tiger data. The project was primarily built through agentic AI. I read the code
Challenges we ran into
We ran into a few issues while building our project. Number one was trying to find a good API for financial data, as a lot of real time ones are pricey. We landed on Finnhub, which has reasonable API limits for the demo. This would however need updating in a production environment
Accomplishments that we're proud of
We are proud of how well the app picks relevant stocks and lets the cards influence the trading decision it creates. The response speed is also something to highlight, as this required a bit of re-architecting to ensure the user wasn't waiting for ages.
What we learned
What's next for Tarot Trader
Built With
- fastapi
- gemini
- postgresql
- react
- tigerdata
- typescript
- vite