Python for Global Applications: teaching scientific Python in context to law and diplomacy students
Anna Haensch
Karin Knudson
For students across domains and disciplines, the message has been communicated loud and clear: data skills are an essential qualification for today’s job market. This includes not only the traditional introductory stats coursework but also machine learning, artificial intelligence, and programming in Python or R. Consequently, there has been significant student-initiated demand for data analytic and computational skills sometimes with very clear objectives in mind, and other times guided by a vague sense of “the work I want to do will require this.” Now we have options. If we train students using “black box” algorithms without attending to the technical choices involved, then we run the risk of unleashing practitioners who might do more harm than good. On the other hand, courses that completely unpack the “black box” can be so steeped in theory that the barrier to entry becomes too high for students from social science and policy backgrounds, thereby excluding critical voices. In sum, both of these options lead to a pitfall that has gained significant media attention over recent years: the harms caused by algorithms that are implemented without sufficient attention to human context. In this paper, we - two mathematicians turned data scientists - present a framework for teaching introductory data science skills in a highly contextualized and domain flexible environment. We will present example course outlines at the semester, weekly, and daily level, and share materials that we think hold promise.
computational social science, public policy, data science, teaching with Python
DOI10.25080/majora-212e5952-00b