Transdisciplinary teaching practices for data science education: A comprehensive framework for integrating disciplines

Abstract

Teaching data science programmes poses challenges for instructors due to the transdisciplinarity of the field and the diverse backgrounds and skill levels of students. Effective data science education requires a comprehensive approach that incorporates theoretical knowledge, practical skills, and industry relevance. However, it is difficult to find appropriate teaching strategies and tools that successfully integrate all these elements into the classroom. Consequently, there is a need to identify and develop effective pedagogical methods, instructional resources, and technological solutions that enable instructors to deliver well-rounded data science education that caters to the diverse needs of students and prepares them for real-world data-driven challenges. Knowing which technology is appropriate to use in conjunction with a particular teaching pedagogy to deliver a particular piece of learning material to diverse students is crucial. Therefore, this study aimed to explore how the TPACK (technological pedagogical content knowledge) influences data science teaching practices. To achieve this, the study surveyed 26 data science instructors to assess their confidence in the seven TPACK constructs. The findings of the study showed a low representation of women in data science education. The findings also showed a balanced knowledge between pedagogy and technological content, indicating that instructors can contribute to a comprehensive and engaging learning environment that supports student success in data science education. Despite this positive finding being established, it was not clear which technological teaching and learning tools instructors are familiar with. To this end, future studies are recommended in this area. The results further showed that model evaluation is not taught at undergraduate level. Therefore, the study recommends continuous professional development for data science instructors to effectively contribute towards training current and future data scientists. This is necessary since technologies, data, and data science tools and techniques evolve. Furthermore, the study recommends research be conducted on the type of data science framework required to guide instructors in terms of curriculum design, pedagogies, and technological tools. Research that informs policy is also necessary to support efforts directed at data literacy, especially to support personnel involved in human capacity development in data science. Lastly, within the scope of data science, interdisciplinary collaboration at national and international levels is recommended so that instructors can stay updated with advancements in subject matter, technology, and pedagogy.