Exploring data landscapes and so much more
It was awesome to be one of the opening speakers at this year’s US Conference on Teaching Statistics this morning at 4am – thank you again Allan Rossman and Kelly McConville for the invitation! Each of the speakers had five minutes to share something related to the conference theme of Broadening horizons, and all of the other speakers in this session gave amazing talks. You’ll soon be able to check out videos of their talks and slides here, which I would highly recommend. In the meantime, you can find my slides here: bit.ly/datalandscapes
[PS it’s not too late to register and attend USCOTS virtually. There are still a bunch of awesome talks and “breakout sessions” on offer over the rest of this week, including one co-presented by Chris Franklin and Pip Arnold on Friday at 5:15am!]
I also have to thank Allan Rossman for inviting me to write a guest post for his amazing Ask good questions blog last year. Writing the guest post Popularity contest encouraged me to share more about what I’ve been exporing in terms of my statistics and data science teaching, particularly in terms of how to design tasks that can engage and support a wide range of students to learn from data. This led to me co-writing a paper with Chris Wild titled On traversing the data landscape: Introducing APIs to data-science students.
For my five minute talk this morning, I focused on this idea of data landscapes and tasks that I’ve been using in my intro stats course (STATS 100) to allow students to customise their data learning opportunities for weekly investigations or explorations. The basic idea is that students choose at the start of the task what specific data landscape they want to travel through for their learning journey. I provide some structure to the task to guide their journey, and as well as submitting evidence that they reached the destination, they also write a travel review (i..e learning reflection) about their journey.
One of the examples I cover in the talk happens in the first week of STATS 100. We explore time series data, and both the tutorial and lab tasks for the week are based around food prices within New Zealand. The tutorial task asks students to choose a meal that they enjoy with family/friends, find a recipe for the meal, and then predict the cost to make this meal a year from now (among other things, like describing the trend/s etc.). The lab task that follows picks up the statistical ideas from the tutorial task, and guides students to learn about creating effective visualisations of time series data using the programming language R, again with the opportunity for students to select different food items to explore and compare how mean prices change over time.
When I wrote the guest post for Allan and the paper with Chris earlier this year, I would describe my data landscapes adventures as being on the domestic level – in that travel was limited to a few locations close to home. But this year, over the six months of teaching of STATS 100 during both summer school semester and semester one, I went full international-level travel – providing student data customisation for every single weekly tutorial and lab task. I definitely have the “travel bug” and now can’t imagine teaching any other way.
What I look for now in terms of data sources are opportunities for students to follow their own curiousity and for their personal interest and genuine not knowing something to be the motivation for their investigation or exploration. I do unashamedly make it all about them for these weekly tasks – What don’t YOU know? Why are YOU interesting in finding this out? How will YOU use this data source to explore or investigate an answer? What did YOU discover about the world through this task?
That’s not to say we don’t also explore important socially and culturally relevant/provocative data and modelling contexts as part of the course – I haven’t even discussed the weekly readings, the notes, interactive examples and three hours of “lectorials” that also make up the learning for each week! You’ll just have to wait to learn more about these other important aspects of the curriculum and assessment design – I plan on writing more posts about my adventures with teaching STATS 100 over the coming months.
In the meantime, if you’re teaching high school level stats and are interested to learn more about using APIs or other cool data science approaches, then please get in touch via my email: email@example.com. I’m already working with a small group of teachers but there’s always room for more #undercoverdatascience