Introduction
In this blog, I would like to share my experience in subject design, one that is driven by understanding the needs of the student and current industry practice.
My field of teaching and research is in the domain of data mining and cybersecurity. Prior to joining Charles Sturt I had worked as a post-doctoral research fellow on a few industry-related data science projects. I had also previously worked as a course coordinator at the University of South Australia (UniSA).
Upon joining Charles Sturt University (CSU), one of my tasks was to develop a foundational data analytics subject as part of our Graduate Certificate in Applied Data Science course. This subject, ITC575, was to serve as a gateway to the various topics within data science including AI, machine learning, statistics, programming and databases.
In my opinion, the process of subject design and development is really a never-ending dynamic continuum. It is almost fallacious to perceive a quintessential state for a subject, without regard to the audience nor the era within which that very perception is being made. For this piece, permit me to deliberately commit this serious blunder, by placing my subject into such a quintessential state. By so doing, it is my hope that, I will be able to shed light on my own subject design experience and how my strategy may have produced a beautiful outcome.
What were you trying to achieve?
My goal in designing this foundational subject was to give students, who have had little or no prior experience in data analytics, a taste of what data analytics involves, and how they could readily apply what they have studied in their own work places. Of course, each time I would get carried away, there would be a jolt in the back of my head to remind me that this subject was meant to be introductory – the course coordinator certainly made sure of this 🙂 . An important factor was to ensure that this subject served as a gateway for students into the other interesting topics of data science. This meant that I needed to cover at breadth, the various domains in data analytics, but at the same time preserving the prospect of students being able to apply what they have learnt in their workplaces.
What did it look like?
The first question I asked myself was “what is the overarching lesson I want my audience to learn?”
While the learning outcomes are clearly spelt out in the CASIMS document to guide subject development, I found it very useful to summarise all of this into one answer: “I want my audience to have a big data analytics starter pack” – broad enough so they evidently see the scope within big data analytics, yet deep enough for them to evidently see how they can apply their knowledge in their own work places to make an impact.
Such a balance is rather an ambitious one, and not entirely easy to achieve. This is especially so when the subject is expected to cater for a cohort of students who may have no prior data science, computer science nor mathematics background.
How can I make this happen?
So what is the current industry practice? And what does a typical student look like? How do I marry the two to achieve my objective? These are some of the questions that ran through my mind.
Thankfully, having worked on several industry data science projects, and consistently partaking in data science conferences and journals in several roles meant that, the first question on industry practice could be answered relatively easily. But the other two questions were not so easy to address.
It is often said that, to be a teacher, one must be a student first. I dare say that, in the context of good teaching, this statement is not as remarkable as it sounds – the latter (to have been a student) is almost always trivially met by the very virtue of being the former (a teacher). Audaciously speaking, a statement I will find more remarkable is “to be a good teacher, one must first aspire to be the student whom one seeks to teach” – and so I did!
So what does a typical student who wants to study a foundational data analytics subject look like? Well, I started putting together two typical but fictitious personae based on my previous engagement in other data science subjects within the course. By engagement, I mean both verbal and non-verbal feedback from students and data provided by the student administration system. I named my personae Bob and Alice.
Bob is a 45 year old male who has worked in public health for a local government for 20 years and is currently a manager. He lives in regional Australia. Bob wants to upskill himself so that he may identify new ways of analysing their health data – possibly, to set up a data science team. Bob has no IT background though he is very experienced in public health. Bob thinks that a grad cert in this field will be very ideal.
Alice is a 30 year old IT teacher in a secondary school. Alice has a previous background in IT. Alice has a part time job designing websites as a freelancer. Alice finds the google analytics on her websites’ traffic very interesting. She wants to upskill herself by learning more about data science. She would like to be able to integrate data science applications into the websites she builds as well as introduce it at a basic level to her high school students. She thinks that the graduate certificate is the best way forward for her.
How do I deliver the overarching learning outcome to these two individuals?
I think of every lesson I deliver as a story to tell: it must have a lesson, a structure, and it must be engaging for that audience. If each topic in the subject is a story that forms part of a cohesive narrative, then it is expected that when they are all put together, the entire narrative should tell one giant story, with the overarching lesson.
Permit me to exemplify this with one such story. In this particular story, the lesson I was aiming for was an appreciation of the real-life challenges associated with big data analytics – I think this lesson was well received. Anyway, prior to our lecture, I asked my students to read up on facebook’s data usage policy and to discuss some of the challenges in our discussion forum – you can’t blame me for trying to encourage some good old peer-to-peer engagement J . Then, in our lecture, we analysed various big data analytic processes and architectures, where we discussed the possible data analytics process used by facebook. My objective was that, as we discussed the processes and architectures, students would be able to relate to the facebook’s data usage policies that they had read up on. In my anecdotal view, students seemed quite engaged in the discussion based on their prior readings – well… those that actually did the prior reading. Of course, students were not off the hook yet with this lesson. As part of their assessment they were asked to design and analyse potential data analytics processes for their own organisations (or an organisation they aspired to work in).
In this story it is evident, I hope, that the lesson was structured in a way that would be engaging even to an audience with differing backgrounds, and differing goals for studying this foundational subject – needless to say that such assessment marking is usually a wee bit more time consuming J . While we have not yet run a formal subject survey ( at the time of this writing), I have already received some positive informal feedback from students on how this process has gotten them to start thinking about how they could apply this knowledge at their work places – I pray this is not another one of those not-to-be-trusted polls (cf. Brexit, Trump/Clinton, ScoMo/Shorten) where students actually slaughter me in the real survey – well …ask me about it next time you see me.
A wise man once told me that good teaching can be likened to good story telling : they both have a lesson, they have a structure and they are engaging. In this blog, I have highlighted my personal journey of subject design by designing the lessons as stories based on student feedback and industry practice; and how these inform the stories I tell through my subjects.
So what stories do your subjects tell?