Iron Ticker

• 

  Primary Screen

Inspiration

Our inspiration for the project was our inspiration for investing and the stock market along with how traders bought on the stock market with hand signals. Our project incorporates some of the used hand signals to buy and sell materials that relate to "Steam and Steel."

How we built it

This program was built with a steam and steel themed dashboard using Python, Streamlit (with a custom CSS file to edit the website modules) and Plotly to display the candlesticks of each product. FastAPI and Pydantic are used to get hand gestures and hand signals from the user's camera, use Gemini's interpretation of whawt it is seeing which then it is converted to either a "Buy", "Sell", or a "Neutral" action with the quantity of what is being either purchased or sold ("Neutral" is 0). The hand gestures are tracked through MediaPipe which tracks key points on your hand like knuckles, fingertips, and other points on your palm to see the orientation of your hand. The Gemini API is used to create these market shocks which affect the stock and it's direction whether it'll rise or fall along with camera recognition. Lastly, the project was deployed and hosted on Vultr using Docker with Ngnix for HTTPS.

Challenges we ran into

The challenges we ran into were how the hand gestures were recognized. Sometimes a user would hold their hand up to the camera for several seconds and nothing would be registered. We had to fix this as this was a main feature apart of our program

Accomplishments that we're proud of

The accomplishments we were proud of was getting a project done like this in this time frame because it is our first hackathon. With that we were proud to connect physical data like hand gestures through a camera into code which then dictates another function in the actual program. We were proud that we were able to get the camera confident on most hand gestures as long as it's in a reasonable range and that holding a hand gesture does not submit duplicate trades.

What we learned

We learned a lot of things from how to use some APIs and libraries that we used to using Gemini API as apart of our project. Seeing how these different modules work together like the hand recognition and AI was a very cool and helpful experience that improves my knowledge base going into the future.

What's next for IronTicker

The main thing we want to improve is the accuracy and timing of hand gestures it would be nicer if it picked it up faster and quicker as sometimes you have to hold you hand in front of the camera for a few seconds. With that, in actual hand gesture trading there are alot more hand gestures that go beyond buying or selling 5 stocks so that would be added too.

Built With

• css
• docker
• fastapi
• google-gemini-api
• html
• javascript
• mediapipe
• nginx
• plotly
• pydantic
• python
• streamlit
• vultr
• websockets