While studying at the University of Technology in Sydney, I worked with Shoujin Wang on a machine learning project in foreign exchange trading. Shoujin is a professor in the data science department at UTS and has won multiple awards for his projects in agriculture, education, and indeed finance. This project was related to predicting the profitability of traders, rather than the more common project of predicting market trends (which is in itself a pretty overdone and useless trend, but that’s another discussion to have).
We worked for a little over two weeks, focusing our energy solely on the project. Using a dataset from UTS collected from thousands of traders worldwide, we trained neural networks, random forest models, and gradient boosted models on the data to see which gave us the best predictive performance. We documented our findings, wrote a report, and presented our findings to multiple professors from UTS.
As someone with a casual interest in finance (shoutout the Money Stuff newsletter by Matt Levine, my favorite source for all things from that world), it was really fun to dive in in more detail. Learning about how money moves around the world teaches you so much about the way things work, and it was also a great way to improve my machine learning understanding and capabilities. Thanks to this project I was able to jump back into training models at Northeastern my senior year, which led to a lot of my current professional and academic interests.