PROGRAM

We've all put a lot of effort into building expert finder systems—but what else can we do with all that data? At the University of California, San Francisco, we took a fresh look at our UCSF Profiles data and asked a deceptively simple question: is this researcher doing basic science, or clinical/translational work? That question turned out to be more complicated—and more valuable—than we expected.

This talk will walk through how our team combined publication, grant, and other research metadata from UCSF Profiles to explore faculty research directions through a new lens. We tested a range of approaches, from rule-based methods to AI models, and worked closely with internal stakeholders to validate whether our classifications held up.

Along the way, we grappled with the fuzziness of research categories, struggled to resolve disagreements over what defines a researcher, developed strategies to build trust in algorithmic results, and figured out how to apply similar techniques to classify grants—not just people.

We thought the project was about labeling researchers, but it made us rethink how we use the data we already have to tell clearer stories about who's doing what at UCSF—and using those insights to support planning, reporting, recruitment, retention, and research development. If your institution has an expert finder system and you're looking for new ways to unlock its potential, this talk is for you.