Abstract
In this paper we reflect on what we have learned in the past three years creating a novel experiential education program in data science for undergraduates. In this Data Science Internal Internship (DSII), students are paid as employees of the university office by which they are tasked with helping to overcome administrative challenges through the introduction of data scientific solutions. Interns meet weekly with the DSII’s program directors in the computer science department, receiving feedback on their design and implementation strategies, technical support as necessary, and guidance on organizational dynamics they encounter along the way. Similar to traditional internships, this offers students the opportunity to acquire data scientific expertise by working with a mentor in a professional setting. A key difference is that the mentor is the actual supervising client, typically with minimal data science expertise. This drives the uniqueness of the DSII’s learning environment, which includes the guarantee to students that they will have access
to the data they need to address the administrative challenge in question. It also grants students regular and direct contact with the supervising domain knowledge experts. This results in direct benefit to the university while providing students the unusual chance to serve as lead developers tackling high-value problems.
We explore the Internal Internship approach as a proof-of-concept, describing the infrastructure that will help make possible its replicability. The Internal Internship offers undergraduates the chance for demonstrable achievement through its expectations of high motivation and creativity in self-teaching, and of effective application of new technologies. These expectations emerge from real administrative challenges requiring innovative solutions, and through direct and dedicated relationships with respective administrators. Because of the distinctive learning environment it offers, we find that the Internal Internship concept is worth pursuing as a complement to the many new data science programs spawned over the past decade.