The undergraduate years are formative. This is where we lay the foundation for a successful future in analytics.

I am a  firm believer that a novice can only understand Statistics once they have spent time with the theory, practiced it with pen on paper and can dialogue with statistical concepts with confidence. I believe that students should initially work in groups to share knowledge and practice the unfamiliar concepts together in a comfortable space, amongst peers. This learning activity also helps hone their skills to, someday, be meaningful contributors in a team environment. It is, of course, essential that they work independently as well to ensure that they do in fact understand and can apply the concepts on their own. My teaching pedagogy is honed from a selection of techniques that I found most effective when I was a student, but I have adjusted these techniques to suit the background of the different students I teach and to resonate with the current generation of young people. I also believe that your teaching is only as effective as your interest in your students' growth.

Students should have a sound mathematical background when engaging with statistics, but unfortunately this is not always the case. One way I've engaged with this problem is to present mathematical problems in a language that students can relate to. I also revise key mathematical concepts of importance to each section of the work, for example derivatives which are important in the derivation of the moments of a distribution. I would normally supply students with a summary of the necessary mathematical concepts, such as a summary of series expansions necessary in the derivation of moment generating functions, especially when it is important that students revisit these concepts many times throughout the semester (↪ view). Alternatively, I would devote a part of the lecture to mathematics revision, for example integration revision essential to the theory of continuous probability distributions (↪ view). Furthermore I attempt to keep the atmosphere light whenever I have to explain a mathematically complicated result or proof. I find that this results in better participation by the class and a generally pleasant lecture. Refer to e-mails from students as proof of the success of my approach (↪ view).

Every new section of a course is started with a roadmap of where we've came from, where we are now and where we are heading. I believe this creates context and keeps the students focused on the core objectives of the course. An example of a STA221 roadmap can be viewed here. This is followed by a discussion of terminology the students should be familiar with from previous years. If I detect any hesitation from their side I address this with revision homework on these concepts when appropriate (↪ view). The homework is always scored and serves as an indication of the students' commitment to learning and provides insight into the students' progress. I also use the results of the weekly tutorials to highlight problem areas to be revisited.

I strive to keep the class atmosphere light. This is a specific strategy that I implement to increase engagement from an otherwise fairly disconnected audience. Let's face it, mathematics and statistics are not every undergrad's cup of tea! And sometimes they take a course because they have to for degree purposes and not because they want to. But, that doesn't mean that I shouldn't seize the opportunity to "bring them over to our side". I do this by varying my pace, engaging in limited yet necessary banter, displaying earnest enthusiasm and verve, appreciation of their current situation and interests (such as exams or varsity shield) and a jovial attitude. As a stark contrast to this, my students are always well informed of the seriousness, difficulty and importance of working hard and I make it a point of reminding them of this when necessary.

Another part of my teaching philosophy is to be an example to my students of being a professional. I am always early for lectures, I am always prepared for lectures, all documentation shared with students are edited and organised and each consultation is done in a respectful and professional manner. Refer to student comments where a student specifically mentions what this meant to them (↪ view). Because, as much as the formative years are about laying a solid foundation, it is also about soft skill development.

At postgraduate level the foundation is laid so my approach shifts towards turning my students into confident problem solvers. 

It goes without saying that students don't learn everything at undergraduate level. Some topics are simply more advanced and require a certain level of maturity before they can be taught. In that sense, my philosophy remains the same as with undergraduate teaching when students need to learn new concepts: roadmaps to always know where we came from, where we are and where we're heading; additional notes when some concepts require refreshing; and structured assignments to test understanding of new concepts. 

Of course now, more than ever before, in the midst of the fourth industrial revolution, students also need to apply their deep understanding of statistical concepts in appropriate statistical software. In my view, computer application cannot simply replace a "pen-on-paper" understanding of statistical techniques. Students still need to understand which statistical technique should be applied when, how the technique works, they need to understand the differences between analytical techniques applicable to different data types, they need to be able to distinguish between the "right and wrong" of statistical analysis. They need to know that machine learning is statistical learning is statistical modelling. And they need to know this with confidence. I am a true ambassador of the rightful place of statistics, and statisticians, in the era of data science. I agree with the comment that "a data scientist is someone who is better at statistics than any software engineer and better at software engineering than any statistician". Statistics is the foundation of data science, the rest is just technology and domain knowledge. Without sound statistical knowledge, there is no trustworthy data science. But, there is also no getting to grips with the vast amounts of data being generated every second of every day without a computer. And so, while absolutely not diminishing the importance of theoretical understanding, all modules I present at postgraduate level are accompanied by thorough practical demonstration in R/RStudio, Python or SAS. All of these software packages are rated among the top in the IEEE list of top programming languages. 

The goal is to graduate skilled and confident analytics problem solvers. To achieve this, and in contrast to the more structured assignments, students get mini projects to solve and then present their solutions as mini professional reports (↪ view). Not only do they become more confident in problem solving, they also acquire further soft skills in the form of report writing.

This teaching approach has been in place for a short period and I can already see the students shining brighter and exuding more confidence (↪ view).