A major characteristic of data science is its focus on doing. My lessons take the form of autograded worksheets, which are given to the students in advance. The autograder allows students to instantaneously test the validity of a variety of solutions on their own. Live coding the worksheets in class allows students to compare their solutions to mine, and allows students to ask about offshoots to the questions. For an example, see the STAT 545 worksheets and the STAT 201 worksheets.
Another major characteristic of data science is its problem-first perspective (as opposed to a model-first perspective). Whenever possible, I draw inspiration from examples encountered in everyday life along with business needs inspired from my consulting and MDS capstone experience. For example, predicting the total number of wedding RSVPs motivates simulation, and preparing for floods motivates quantile regression. A problem-first focus identifies areas that are not well developed or not often taught in traditional statistics curricula, such as non-Gaussian predictive distributions. Filling these holes form the core of my educational leadership activities.
Giving a course an open access platform through a website allows a wider reach and brings attention to our department and university. Also, open access is like an accountability partner: it provides incentive for producing higher quality resources. Plus, open access frees students from the need to squirrel away resources on their own. Open access gives the course an entity of its own; a community with a clear vision in which to join. For an example, see the STAT 545 course website: https://stat545.stat.ubc.ca/
I give my teaching assistants the opportunity to add an additional line to their CVs by inviting them to deliver guest lectures and to develop course material. Together with bringing TA’s into the decision making process about course delivery and direction, the result is an engaged teaching team and an improved sense of learning and community for the students.
)