Review of “Computational Investing, Part I” taught by Tucker Balch (Fall 2012)

Posted on May 25, 2013 by


ALERT: This is an old review based on the Fall 2012 session. Please follow this link  to see the new review.

This is a review based on survey responses by 2,350 students who enrolled in Fall 2012. There is a new review available.

Related articles


If you are considering to take the course, please take a look at the results of this survey. I believe they may help you decide if the course is right for you. Completing this course will require a significant investment of your time, and I want to be sure it is a good match!

We are using the results of this poll to help us revise the course for the next offering, so some of the issues pointed out by the students will be resolved before the next offering starts.

Satisfaction with the course by students who completed the course

The students who completed the course were asked “How much do you agree with each of these statements regarding your learning in the course? (Please rate on a scale of Strongly Disagree to Strongly Agree).” We report the percentage of students who selected Agree or Strongly Agree below:

  • Considering everything, the instructor was an effective teacher: 61.1%
  • For the amount of time I invested in this course, I’m happy with what I learned: 67.5%
  • The course materials were presented in an engaging manner: 57.0%
  • I would like to take a more advanced course on this topic: 92.2%
  • I found the course personally fulfilling: 59.0%
  • I learned what I was hoping to learn in this course: 42.8%

Satisfaction with the course by students who did not complete the course

  • Did you find the course useful even though you didn’t complete it? Yes (89.7%)

What was the best thing about the course?

These answers were collated from the survey responses by Georgia Tech Ph.D. student Brian Hrolenok. Brian will be one of the TA’s for the Spring session of the course. We asked Brian to collect the 5 most frequent/similar answers to this question.

1. QSTK: The single most frequent response to this question had to do with QSTK (this is the software package we use for analysis in the course). Students thought both that the library itself was useful, and that using QSTK and the software tools taught in the course helped in applying the concepts in practice. Some relevant quotes:

“The QSTK software provides a great example for implementing quantitative trading.”
“QSTK and the ability to apply the concepts directly in code. During the lectures (which were pretty good), I could see conceptually how to write a program. Having the toolkit handy that handed much of the computation (database load, data cleanup, graphing, event analysis, etc) helped me concentrate more on defining strategies and less on gathering data. “
“There is tremendous value in sharing the software architecture of a quant finance system, and I appreciated this unique knowledge from”

2. Python and libraries: Students liked learning Python, and if they already knew the language, liked learning how to use it to do analysis on financial data. Specific libraries that were mentioned included Pandas, CVXOPT, and Numpy. Some relevant quotes:

“Programming in Python with Yahoo data on stocks – very engaging, interesting”
“Building the project was a fun hands on approach to learning the use of python, pandas and Numpy”
“Getting a usable Quant framework, learning Pandas” 

3. The assignments: Many students said they enjoyed the programming assignments, both for the way they were linked together in a course-long project, and as practical applications of the content from the lectures. Some relevant quotes:

“The practical homework and programming assignments based on real approaches.”
“The assignments all helped building one final program, instead of being disconnected”

4. Practical applications: Students said they enjoyed learning about how algorithmic trading is done in practice.Learning about event studies and how hedge funds operate were mentioned specifically.

5. Understanding the stock market: Students also enjoyed learning about the mechanisms behind how the stock market actually works. Hedge funds were mentioned specifically, as well as High Frequency Trading, CAPM, Portfolio theory, and the Sharpe ratio.

From the first session: What needs the most improvement?

These were problems reported by students who took the course in Fall 2012. We addressed many of these issues in the second session, Spring 2013.

1. Needs a more thorough QSTK/Python introduction: Many students who were not familiar with python wanted a more in-depth introduction to the language, as well as getting the development environment up and running. Students had a number of suggestions ranging from making python a hard prerequisite for the course to providing a supplementary optional set of lectures for introducing python, the libraries, and QSTK. One student mentioned a Wes McKinney pandas talk, here’s a link to a video from his pycon 2012 talk:

Instructor response: We restructured the course to add some introductory Python material at the beginning. And the students in the second session seemed to benefit from that. We focus on Numpy/Scipy data manipulation and a little bit on Pandas. However, students should be aware that we do expect strong programming skills as a prerequisite.

2. Needs to be better organized: Many students thought the course as a whole should have been better organized. Delays in assignments and video lectures were the most frequently mentioned problems, but students also suggested having lecture notes to accompany the videos, improved forum support, and an improved wiki. One student suggested following the example given by another Coursera course called “Think Again”:

Instructor response: We rearranged the order of some of the videos and added others to fill in gaps.  Lots of students who took the first session and the second mentioned that they noticed that the course was better organized.

3. The course content is too basic: Some students wanted the course to be more in-depth, but almost all of them disagreed about which parts warranted it. Equal numbers wanted a more detail on the programming and less on the finance as vice versa. A number of students wanted less time spent on the programming aspects, contrasting directly with those that thought not enough time was given to introducing Python and the development environment. Several students suggested having links to additional resources for more advanced topics.

Instructor response: We’re going to keep the topic coverage and content mostly the same. We think it is better for students to find other courses that are a better match for what they want to learn. Please take a look at the syllabus to see if the topics are appropriate for you.

4. Needs better production quality for the video lectures: A few students pointed out that at times the writing being displayed or gestures to things that had been written were not part of the video:

“Once the format of the videos changed partway through the course, we were no longer able to see what the professor was drawing/writing in real-time. We only got to see the finished sketch. It is helpful to be able to see what he is drawing while he is drawing it as his explanation corresponds with what he is drawing.”

In general, students felt that some of the video lectures looked rushed.

Instructor response: That is very interesting. I did not realize that actually seeing the drawing happening was important.  When we reshoot these sessions, we’ll work to address this.

5. Needs to focus on more practical content: Several students mentioned practicality, however there were at least an equal number of compliments on linking the programming assignments and material to how things are done in practice. Some students seemed to be looking more for a guide to investing than a computational approach. Relevant quotes:

“Hope the instructor can provides more information on practical trading strategies, algorithms and implementation. “
“practical investment tips”
“More practical problems and examples such that we can use in real trading.”

Instructor response: As one response to this, we travelled to koh samui luxury resorts and explored one of the major US Exchanges, the CBOE.

Participation in online forums

We found that participation in the forums was a strong predictor of success. Of students who completed the course 95.8%  read the forums.  Of those who did not complete the course, only 62.4% read the forums.


Fatima Wirth, Ph.D. assisted with instructional design, and the course was TA-ed by Sourabh Bajaj.