Applied Business Analytics from MIT
A review of the 6 week executive program that newbies can use to enter the world of data
While randomly browsing through my Instagram feed last May, I came across an advertisement about the ‘Applied Business Analytics Executive Program’ from MIT’s Sloan School of Management. I am not sure if it was the course’s name or the school’s name that caught my attention (I almost never stop to look at ads) but I clicked on the link and checked out more information from MIT’s website.
The program is a 6 week certification program that covers, through case studies, the concepts of regression, decision trees, random-forests and clustering for $2800. After two days of contemplation and consulting with my data scientist wife, I enrolled into the program. The 6 week program was quite an eye opener into the field of data analytics for a newbie like me.
In this article I hope to summarize the program’s content and list what I liked about it and where I think the program can be improved.
Who is this program for?
If you are like me and want to understand the fundamentals of data science including the meaning and applications of terms like regression, decision trees, random forests and clustering then this program is definitely a good way to start. As MIT’s brochure says, this program is for “..anyone who wants to understand the business applications of analytics for functional practice or general management”. It is primarily planned for non-technical professionals (you will not be doing any programming) although snippets of the code behind the examples are provided for the programmers in us. They show these snippets in R, Python and Julia (an MIT created programming language that is super-fast compared to Python).
What will you learn?
Clustering:
Using Netflix’s movie recommendation feature as an example, the program introduces the concept of clustering. The course introduces the types of clustering (K-style and hierarchical) and describes in detail how Netflix uses clustering to provide recommendation on movies that the user will like.
Regression:
The program explains both linear and logistic regression concepts through two case studies. The famous Moneyball story is used to discuss how linear regression establishes a relationship between a dependent variable and one or more independent variables. The Farmingham Heart Study that is being done to understand cardiovascular disease is the subject to discuss logistics regression. Here the program introduces the concept of training and validation data sets.
Decision Trees and Random Forests:
Two case studies are used to introduce the concepts of Decision Trees and Random Forests. Using decision trees, the cost of housing in Boston is evaluated based on the user’s preferences. This concept is extended further in Random Forests where the program touches on a project that predicts which way a supreme court decision is likely to go based on historical data. This is then reiterated through two other case studies of improving inventory management (decision trees) and maximizing investment portfolio value (random forests).
Machine Learning:
A good chunk of the program focuses on case studies on various machine learning concepts. Using different examples, the program introduces concepts of natural language processing (NLP) and predictive modeling.
What I liked about this program:
For someone new to the concepts and nuances of data science like me, this program gave an excellent introduction and a platform from which one can learn further and build their knowledge. As I was going through the program, I was able to understand how I can apply some of these techniques in my job and how it can help my work in the field of reliability prediction and prognostic health monitoring. While this program isn’t a magic wand that would make you a data scientist in 6 weeks, it trains you to look for opportunities to use data towards a meaningful end. How you use this knowledge and build upon it is up to you and your curiosity bug.
Opportunities for improvement:
To put it bluntly, I felt the program was too easy. Perhaps it was my expectation that a program from MIT would require rigorous studying and tough assignments that left me surprised and wanting to do more each week. I felt that the assignments that were given were rather too simplistic and the lack of a serious capstone project (they ask for a document where you can talk about how you will use the concept you have learned but you are not submitting a completed project) made me feel that I finished the program rather too easily and I didn’t earn it.
Conclusion:
Overall, I would definitely recommend this program to those who are looking to venture into the field of data analytics. Through well-chosen examples delivered by very knowledgeable faculty, the program beautifully introduces key concepts of analytics that business leaders and new project managers can use to drive critical business decisions meaningfully or manage data science projects effectively.
