Harvard Data Science Review’s Founding Editor-in-Chief, Xiao-Li Meng, and Media Feature Editor, Liberty Vittert Capito, interviewed Eric LeVine, founder of CellarTracker, the world’s earliest and largest crowd-sourced wine cellar and review platform, about what wine can teach us about data, AI, and human behavior. The conversation starts with how LeVine, a former Microsoft program manager, turned a personal spreadsheet for cataloging a single cellar into a massive community-driven database revealing how people buy, drink, rate, and talk about wine. They discussed the subjectivity of taste and the challenges this poses for building recommendation engines, how CellarTracker is experimenting with “digital twin”–style recommender systems and AI-generated wine summaries, and what it takes to keep such a large, complex, and highly personal data set usable and trustworthy—from data quality and label recognition to privacy-by-design and a strong ethical principle of avoiding intrusive uses of user data. LeVine also shared his vision for how AI could reshape wine discovery and commerce over the next decade, from restaurant list guidance to more consumer-centric retail, and reflected on what personalized wine prediction can teach us about personalization in other, more complex domains.
LeVine et al. (Fri,) studied this question.