Reflecting on the past few months
At the end of 2019, during the short break between the Fall and Winter semesters of the Rotman MMA program, I had the chance to reflect upon what I learned in the past few months. During the orientation, our class was notified that the program would get very intense very quickly, but only now do we fully understand why. The reason is in the name of the program, which hints at its duality. Management analytics is the intersection of two vast fields of study: management science and data science. It is important to develop knowledge and skills in both these fields to learn what kind of analytics projects certain businesses should or should not pursue.
Management science meets data science
While these two fields are vast and useful independently, to help large organizations benefit from AI in the modern economy, knowledge and skills from both fields must be leveraged. The reason it is important to study both these fields simultaneously is to learn about the challenges one must be concerned with to be successful in a management analytics career. The Rotman MMA program has hosted many events, featuring panels and keynote speakers with many years of experience in these fields. They have placed emphasis on the needs to understand unique challenges from both business management and big data analytics. Technical skills must be complemented by business acumen and soft skills, and vice versa, in order to be able to effectively assess the potential risks and rewards of machine learning and AI projects.
Balancing business management and data analytics
The courses offered by the Rotman MMA program seek to balance the focus on both these. Throughout the program, various short colloquium modules teach students about important technical concepts such as Python programming, AI ethics, data acquisition through API and web-scraping, and social network analytics. These skills, which along with learnings from the full-time courses, are then applied during the practicum project which students spend at least one day a week working on. The fall semester included four full-time courses. Two of these courses focused on the business applications of data analytics though visualization and model development, while the other two focused on potentials and limitations of various machine learning methods.
Studying so many topics at once can be quite daunting, but early in the program, I was given a great piece of advice during a coffee chat I had with a data analytics professional from a major Canadian bank.
Pick specific applications of machine learning and AI of interest, and focus my studying on developing the knowledge and skills that would help me develop such applications.
This was very useful advice that I would encourage anyone interested in or currently studying management analytics consider. Such clarity can help one to focus their efforts on ensuring they are ready to implement these techniques effectively at large organizations.
Understanding models in a financial context
In my case, as someone aspiring to apply data analytics in the financial services sector, I have learned that despite their effectiveness, unsupervised deep learning models are not commonly utilized. The reason is that even if they can possibly result in better accuracy than supervised learning methods, the poor interpretability of the outputs from unsupervised deep learning models limit how actionable the results are for the firm. For example, banks are hesitant to leverage “black box” models for fraud detection because it provides no basis for which to justify action actions which might be taken in response to cases flagged as potential fraud.
The Master of Management Analytics is designed to give students the advanced data management, analytics, and communication skills needed to become an analytics professional.