Building College Grant Capacity for Project Vision
- Impact Allies was hired by Project Vision, an NSF ATE Mentoring organization, to use big datasets and AI modeling to find colleges ready for funding but without a strong grant culture
- Beforehand, PV was having trouble recruiting the right colleges into its mentoring cohorts
- Results were published in J-ATE article
- Details of Project are below
This work collected and reviewed large scale data from multiple sources like the community college scorecard, ASPEN institute, and Achieving the Dream on every community college. All told, 3,000 unique data points were collected per college. PV then developed a cluster algorithm that reduced the complexity of 3,000 data points per institution into two principal dimensions that could be visualized for each college, Figure Right. This method allowed us to identify community colleges that looked statistically like those that had successfully received ATE funding but that had not yet entered the NSF ecosystem. PV applied this approach to find colleges with hidden potential for cultivating a sustainable grant culture. For example, Northern Idaho College and Great Basin College, both colleges without NSF funding, were flagged by the algorithm as resembling funded peers, and subsequently became mentee institutions in Cohort 5. This data-driven targeting not only maximized the efficiency of recruitment but also ensured that Project Vision invested mentoring resources where the probability of long-term institutional change was the highest.
