Automated Stock Trading Bot
T.E. Truong, J. Junyan, A. Gibbons
Hackathon 2018
About
I led a team of three students that designed and implemented an automated stock market trading bot that traded live using real currency for Bitcoin stock. This project was created for the Hack The U competition 2018, the largest hackathon event in utah. My team won first place and we were featured on the department website, here.
The project was split between three parts. The first, created a framework to decide which algorithmic trading model to use as well as making buy/sell decisions through Robinhood. The second and third parts, created the algorithmic models (Bollinger Bands) and the second model (machine learning).
As part of the demonstration, the framework was deployed onto the stock market where it bought and sold portions of real Bitcoin autonomously. Based on profit earned during the 2-hour demo time with initial investment of $1, the team calculated that they could raise one million dollars in approximately 58,000 years.
During my spare time, I try to come up with various models based on concepts I learned in class. One of my attempts fitted a mass spring damper model to large fluctuations of price data with moderate success. Other attempts have used concepts in Physics, Statistics, and System Dynamics.
An Image of the live output graph. Model 1 is the Bollinger Band method, Model 2 is the Machine Learning method. Red is buy, green is sell.