Statistics professor Johan Ferreira was feeling overwhelmed by the amount of “screen time” involved in online learning in 2021. He imagined students must be feeling the same way, and wondered what he could do to inspire them and make his subject matter more appealing.
One of the topics in statistics is time series analysis: statistical methods to understand trend behaviour in data which is measured over time. There are lots of examples in daily life, from rainfall records to changes in commodity prices, import or exports, or temperature.
Ferreira asked his students to write a short, fictional “bedtime” story using “characters” from time series analysis. The results were collected into a book that is freely available. He tells us more about it.
Why use storytelling to learn about statistics?
I’m fortunate to be something of a creative myself, being a professional oboe player with the Johannesburg Philharmonic Orchestra. It’s a valuable outlet for self-expression. I reflected on what other activity could inspire creativity without compromising the essence of statistical thinking that was required in this particular course I was teaching.
I invited my third-year science and commerce students at the University of Pretoria to take part in a voluntary storytelling exercise, using key concepts in time series analysis as characters. Students got some guidelines but were free to be creative. My colleague and co-editor, Dr Seite Makgai, and I then read, commented on and edited the stories and put them together into an anthology.
Students gave their consent that their stories could be used for research purposes and might be published. Out of a class of over 200 students, over 30 contributions were received; 23 students permitted their work to be included in this volume.
We curated submissions into two sections (Part I: Fables and Fairy Tales and Part II: Fantasy and Sci-Fi) based on the general style and gist of the work.
The project aimed to develop a new teaching resource, inspire students to take ownership of their learning in a creative way, and support them through informal, project-based peer learning.
This collection is written by students, for students. They used personal and cultural contexts relevant to their background and environment to create content that has a solid background in their direct academic interests. And the stories are available without a paywall!
What are some of the characters and stories?
Student Lebogang Malebati wrote Stationaryville and the Two Brothers, a tale about AR(1) and AR(2). In statistics, AR refers to processes in which numerical values are based on past values. The brothers “were both born with special powers, powers that could make them stationary…” and could trick an evil wizard.
David Dodkins wrote Zt and the Shadow-spawn. In this story, Zt (common notation in time series analysis) has a magic amulet that reveals his character growth through a sequence of models and shows the hero’s victory in the face of adversity. He is a function of those that came before him (through an AR process).
Then there’s Nelis Daniels’ story about a shepherd plagued by a wolf called Arma (autoregressive moving average) which kept making sheep disappear.
And Dikelede Rose Motseleng’s modern fable about the love-hate relationship between AR(1) (“more of a linear guy” with a bad habit of predicting the future based on the past) and MA(1), “the type of girl who would always provide you with stationarity (stability).”
What was the impact of the project?
It was a deeply enriching experience for us to see how students see statistics in a context beyond that of the classroom, especially in cases where students reformulated their stories within their own cultural identities or niche interests.
Three particular main impacts stand out for us:
students have a new additional reference and learning resource for the course content
new students can refer to the experiences and contextualisation of this content of former students, leading to informal peer learning
students engage in a cognitive skill (higher-order and creative thinking) that is not frequently considered and included in this field and at this level.
In 2024, shortly after the book was published, we asked students in the time series analysis course of that year to read any one of four stories (related to concepts that were already covered in the course material at that point in time). We asked them to complete a short and informal survey to gauge their experience and insights regarding the potential of this book as a learning resource for them.
The 53 responses we got indicated that most students saw the book as a useful contribution to their learning experience in time series analysis.
One positive comment from a student was:
I will always remember that the Random Walk is indeed not stationary but White Noise is. I already knew it, but now I won’t forget it.
Will you build on this in future?
It is definitely valuable to consider similar projects in other branches of statistics, but also, in other disciplines entirely, to develop content by students, for students.
At this stage, we’re having the stories and book translated into languages beyond English. In large classes that are essential to data science (such as statistics and mathematics), many different home languages may be spoken. Students often have to learn in their second, third, or even fourth language. So, this project is proving valuable in making advanced statistical concepts tactile and “at home” via translations.
Our publisher recently let us know that the Setswana translation is complete, with the Sepedi and Afrikaans translations following soon. To our knowledge, it’ll be the first such project not only in the discipline of statistics, but in four of the official languages in South Africa.
Johan Ferreira receives funding from the Centre of Excellence in Mathematical and Statistical Sciences, based at the University of the Witwatersrand, towards this Scholarship of Learning and Teaching project.
By Johan Ferreira, Professor, University of the Witwatersrand
