Temporality in data science education: Early results from a grounded theory study of an NSF-funded cybertraining workshop

Abstract

Interest in data science, especially within the context of graduate education, is exploding. In this study we present initial results from an ongoing qualitative study of an interdisciplinary cyberinfrastructure-focused NSF-funded graduate data science education workshop hosted at an iSchool in the US. The complexity of the workshop curriculum, the participants’ and instructors’ disparate disciplinary backgrounds, and the technical tools employed are particularly suited to qualitative methods which can synthesize all of these aspects from rich observational, ethnographic, and trace data collected as part of the authors’ role on the grant’s qualitative evaluation team. The success of the workshop in equipping participants to do reproducible computational science was in part due to the successful acculturation process, whereby participants comprehended, altered, and enacted new norms amongst themselves. At the same time, we observed potential challenges for data science instruction resulting from the rhetorical framing of the technologies as inescapably new. This language, which mirrors that of a successful grant proposal, tends to obscure the deeply embedded and contingent history of the command-line technologies required to preform computational science, many of which are decades old. We conclude by describing our ongoing work, future theoretical sampling plans from this and future data, and the contributions that our findings can provide to graduate data science curriculum development and pedagogy.