Quant Finance Project Ideas
Whether you are a student, a mid-career switcher, or interested in kinds of projects to build your portfolio, here's some advice on what projects to work on.
Whenever someone asks me how to get into a field involving programming or machine learning, the advice is always to "do projects and upload them to GitHub."
The same applies to quantitative finance, but what quantitative projects would impress a prospective employer?
Ph.D. Theses
A Ph.D. Thesis is one of the final pieces of academic assessment a student goes through before getting their Doctorate degree in their field of study.
The main aim is to show that the person can study a new idea on their own and explain the findings in a way that's easy to understand.
In an article, What’s the point of the PhD thesis?1, (Gould, J. What’s the point of the PhD thesis?. Nature 535, 26–28 (2016)), Shirley Tilghman, a molecular biologist and former president of Princeton University in New Jersey, says she :
sees merit in the monograph form of the thesis. It demonstrates scholarly ability by requiring students to “frame the historical context of a problem, describe in detail the purpose and execution, and then come to a credible conclusion.”
This gives us insight into how we can create a framework for Quant Finance Project Ideas
Provide the historical context of a problem
Describe in detail the purpose
Describe in detail the execution
Come to a credible conclusion
Finance, Economics, Business Ph.D. Programs
There are Finance, Economics, and Business Ph.D. programs at almost all large research universities (some may have a mixture of these).
This means these programs will have had students who wrote a Ph.D. thesis.
This means that there should be a list of Ph.D. theses for each program that we can find.
Quant Finance Project Idea: Replicate part of a Thesis
What quantitative projects would impress a prospective employer?
The replication of part of a Ph.D. thesis works wonderfully based on the framework from above:
Provide the historical context of the quantitative finance problem
Describe in detail the purpose of your work
Describe in detail the execution (data, programming, math, modeling, algos)
Come to a credible conclusion (show you were able to replicate the thesis)
This write-up then gives you a sense of a historical problem and how someone solved it, and since you already know the answer (they wrote it), you can rest easy in that you won’t get too lost!
Of course, if you just copy and paste their work, you won’t understand it, and if an interviewer asks you about it and you can’t say anything about it, then you fail. But that won’t happen because you care.
An example
The University of Chicago is known for its economics and finance departments (amongst other things), so let’s start there.
I Google “chicago finance phd thesis”.
Scrolling down the list of results I see “Access to UChicago Dissertations”
I click on this and see
I click on the Knowledge@UChicago link and then type finance in the search box
Which gives this result:
Without looking into each document just by title alone, we’ll go with the 5th one, “Machine Learning Models Applied To High Frequency Financial Data”.
So we click that one and arrive at
Great!
Let’s look at the first few sentences of the abstract (I left the main idea of each chapter and removed the rest of the description):
Chapter 1 is titled "A dynamic network model for high frequency order flows in financial markets." This chapter constructs a network model for high frequency trading volume data using a regularized vector autoregression moving average (VARMA) method……….Chapter 2 is titled "How useful are machine learning tools in predicting high frequency returns?" This chapter explores the predictability of high frequency returns for 63 exchange traded funds by applying recently popularized machine learning techniques. I consider factor models, LASSO, and random forests….
This looks good, let’s check out the PDF (scroll down the page) to look at the table of contents:
Bingo - we’ll learn about the data, the methodology, some models, what results happen, and their conclusion.
We can reference then thesis when we write up our project and slowly work through understanding what they did.
Personal Example
A long time ago, I was a Statistical Arbitrage Group summer intern at a hedge fund you’ve heard of.
Part of my work that summer was replicating a 1995 Finance Ph.D. thesis at MIT titled Essays on Risk and Return in the Stock Market2. The project was replicating Chapter 1 of the thesis, which focused on Improving Covariance Matrix Estimation for large stock portfolios:
It took most of the summer and involved reading the paper, looking up everything I didn’t know, getting the data, presenting why it was important, writing up some Matlab code, showing results, and proving that I had understood it well enough to implement.
How to choose a project
How, then, do you choose a project? Start looking at different departments and their thesis to see if something catches your eye. The idea is to work on something related to quantitative finance that appeals to you so you will see the project through.
Good luck!
That’s all for today :) For more Quant Finance treats, check out our archives.
Stay quanty!
All the best,
Sebastian Gutierrez
https://doi.org/10.1038/535026a
https://dspace.mit.edu/handle/1721.1/11875