Online Program

Tales from the Trenches in Academic/Industry Predictive Modeling Partnerships (304155)

Keywords: Predictive Modeling, Lessons Learned, Academic Partnerships

The lack of qualified data scientists to build predictive models leaves many companies turning to academia for their modeling needs. Partnerships between academia and industry have many benefits including experiential learning for student and access to the latest techniques to gain a competitive advantage. Often the partners on both sides of the fence jump in with high expectations for successful projects only to hit some unexpected road blocks and delays. This talk will detail the lessons learned in predictive modeling in the context of three projects with external partners to the Virginia Tech Statistical Applications and Innovations Group. Key points discussed will be of interest to those in the areas of both industry and academia allowing them to be better prepared to undertake future projects. Attendees will gain insights into working with administrative data, educating partners, preparing data, implementing models, working around the academic cycle, planning work with students, and much more. The case studies include models that predict employment retention, safety violations, and university engagement.