Home ยป Blog ยป Computational Investing

Computational Investing

100% score on Computational Investing, Part 1
100% score on Computational Investing, Part 1

About a month ago or so I wrapped up the “Computational Investing, Part 1″ course at Coursera.org. A few days ago the certificates have been released, so I figured, I write a post showing of my achievement and at the same time share my thoughts on the course.

As the course name implies it is about using computation / computers to go about making investment decisions. The focus is on investing in stocks, but the principals taught in the course easily transfer to other equity types, like Mutual Funds, ETFs, Bonds, Foreign Currencies and maybe even to futures and options. The instructor of the course is Tucker Balch of the Georgia Institute of Technology. Tucker is truly passionate about the field and what I like most is his practical approach to the subject. All computational aspects of the course are done in the Python programming language and there is a Toolkit known as the QuantSoftware Toolkit aka QSTK that will do most of the heavy lifting in the course. Still, I think that you need to have above beginner level Python skills, be familiar with libraries like Pandas, Numpy and matplotlib and have a basic understanding of Statistics to be able to successfully finish this course. If you have this skill level, then the toolkit will really help on being focused on learning and understanding the fundamentals of computational investing without getting lost in the little programming / implementation details. Having said that, I do not think that I will use the QSTK to implement my own strategies, but it is a great learning and strategy discovery toolkit. In all fairness the toolkit has many more features, like machine learning, that I have not discovered (yet)  and that are not covered by the course.

The course covers the following theoretical subjects: Evaluating and Optimizing performance of a portfolio, Sharpe ratio, Market Mechanics, the Capital Asset Pricing Model, the Efficient Market Hypothesis.

Finally, in the course you will set up a backtest framework that uses event studies to generate buy en sell orders and allows for reporting of and optimizing the performance. The purpose of this is mainly to get familiar with the practice of backtesting, learning the nots and bolts and pitfalls like survivor bias, look ahead bias, over optimization, etc. along the way.

In conclusion: For me the course was a great fit to my interests and skills. I gained a better understanding of how the financial markets work and how I can go about backtesting strategies.

Here is a promo to the course by Tucker.


And here is an intro that talks more about the course in detail.


The course is also offered as a signature track, which means that all of your accomplishments are verified so they can be attributed to you without any doubt. I haven’t opt for this as my main interest is gaining knowledge to use trading my own account and less so recognition of accomplishments.

Part 2
This course is noted to be Part 1. So what about part 2? Well, there is a preview to part 2 of this course below. Part 2 is about how Machine Learning can help in investing. I would be very interested in taking part 2 of this course and hope it will become available on line soon. Currently part 2 of this course is not being offered on line, but only on site at Georgia Tech according to the info on the QSTK Wiki.