Final project in the graduate course "Mathematical Methods - Financial Price Analysis" at Columbia University
-
For this project, as part of a team, I have developed and coded an automatic trading algorithm that, guided by statistical tests (variance ratio and push & response) analyzed actual high-frequency data on prices of certain commodity futures.
-
I coded this in Mathematica and optimized it so that the analysis of 30 years worth of data with 5 min resolution can be run on a personal computer.
-
Based on these analyses and using the "trend following" behavior, the code then simulated real-time trading, with the goal of maximizing the profit while minimizing drawdown.
-
My project ranked first in the class.
-
The interactive demonstration of the code is available on GitHub Pages. However, Wolfram's CDF plugin is needed, which is currently not supported on Chrome browsers.