Resumen del episodio
This latest episode of the Harvard Data Science Review Podcast takes a closer look at election polling: how polls are conducted, what they can tell us about voters, and why measuring public opinion has become increasingly challenging. In advance of the Harvard Political Analytics 2026 conference, hosts Liberty Vittert Capito and Xiao-Li Meng are joined by two conference speakers: Anthony Salvanto, CBS News’s director of elections and surveys, and Michael A. Bailey, author of Polling at a Crossroads: Rethinking Modern Survey Research. From declining response rates and changing turnout patterns to weighting, online panels, and election-night projections, Salvanto and Bailey explore what goes on behind the numbers—and why modern pollsters are increasingly also modelers. They discuss the assumptions and human judgment embedded in polling, the importance of data quality, and the challenge of measuring and communicating uncertainty when traditional measures such as the margin of error tell only part of the story.…
This latest episode of the Harvard Data Science Review Podcast takes a closer look at election polling: how polls are conducted, what they can tell us about voters, and why measuring public opinion has become increasingly challenging. In advance of the Harvard Political Analytics 2026 conference, hosts Liberty Vittert Capito and Xiao-Li Meng are joined by two conference speakers: Anthony Salvanto, CBS News’s director of elections and surveys, and Michael A. Bailey, author of Polling at a Crossroads: Rethinking Modern Survey Research. From declining response rates and changing turnout patterns to weighting, online panels, and election-night projections, Salvanto and Bailey explore what goes on behind the numbers—and why modern pollsters are increasingly also modelers. They discuss the assumptions and human judgment embedded in polling, the importance of data quality, and the challenge of measuring and communicating uncertainty when traditional measures such as the margin of error tell only part of the story.…
Episodios
From Polling to Prediction Markets: How Reliable Can We Predict Elections?
EscuchadoThis latest episode of the Harvard Data Science Review Podcast takes a closer look at election polling: how polls are conducted, what they can tell us about voters, and why measuring public opinion has become increasingly challenging. In advance of the Harvard Political Analytics 2026 conference, hosts Liberty Vittert Capito and Xiao-Li Meng are joined by two conference speakers: Anthony Salvanto, CBS News’s director of elections and surveys, and Michael A. Bailey, author of Polling at a Crossroads: Rethinking Modern Survey Research. From declining response rates and changing turnout patterns to weighting, online panels, and election-night projections, Salvanto and Bailey explore what goes on behind the numbers—and why modern pollsters are increasingly also modelers. They discuss the assumptions and human judgment embedded in polling, the importance of data quality, and the challenge of measuring and communicating uncertainty when traditional measures such as the margin of error tell only part of the story.…
The Impact of AI on Podcasts: From Linear Digression to SuperDataScience
EscuchadoThis month’s episode of the Harvard Data Science Review Podcast turns the microphone on two people who regularly bring data science, machine learning, and AI to podcast audiences: Katie Malone, host of Linear Digressions, and Jon Krohn, host of SuperDataScience and co-founder and CEO of AI software company Y Carrot. They join us to explore how AI is transforming not only what podcasters talk about, but how podcasts are researched, produced, and shared. From using AI as a research partner and production “sidecar” to generating episode summaries, newsletters, and animated video, Malone and Krohn share where AI adds real value and where they deliberately keep humans in control. The conversation tackles questions of authenticity, transparency, authorship, and trust, as well as the potential of AI-generated podcasts as personalized tools for learning. Looking ahead, they consider how AI may reshape podcasting itself: making production easier and more powerful while raising new questions about creativity, human…
Active Industrial Learning: What We've Learned—and What We'd Like to Learn From You
EscuchadoThis month’s episode of the Harvard Data Science Review Podcast takes listeners behind the scenes of Active Industrial Learning, HDSR’s column exploring how data science and AI are applied in real organizations. We speak with column co-editors Hamit Hamutcu and Miguel Paredes about the challenges of translating data science theory into practical business impact. Drawing on their experiences working with industry leaders, they discuss data and AI literacy, responsible AI, organizational transformation, and the critical role of leadership in successful AI initiatives. The conversation also explores the future of enterprise AI, the value of cross-disciplinary collaboration, and how Active Industrial Learning is evolving to showcase lessons from business, education, the arts, and beyond. Whether you're leading AI initiatives, building data-driven organizations, or simply interested in how AI succeeds in practice, this episode offers valuable insights into the people, processes, and perspectives shaping the…
Recreations in Randomness: From Glicko Rating to World Cup
EscuchadoThis month’s episode of the Harvard Data Science Review Podcast explores the rapidly evolving world of sports analytics and how advances in data science are transforming the way we understand competition. We are joined by Harvard statistician Mark Glickman, creator of the Glicko rating system, and sports statistician Stephanie Kovalchik to discuss the technologies, models, and data driving modern sports. From real-time player tracking and probabilistic rating systems to AI-assisted coaching and predictive modeling, the conversation examines how statistical methods continue to shape decision-making on and off the field. Glickman and Kovalchik also explore why traditional statistical models remain central to sports analytics, how access to high-quality data continues to limit innovation, and what emerging AI tools may—and may not—bring to the future of the field. The episode concludes with a look at Recreations in Randomness, HDSR’s column on the many ways data science enriches our recreational lives, and an…
The Judgment of Paris at 50: Wine, Wisdom, and What We Still Don’t Know
EscuchadoThis month’s episode of the Harvard Data Science Review Podcast uncorks the fascinating intersection of wine, judgment, and data science. Economist and wine expert Orley Ashenfelter and Master of Wine Susan Lin join us to explore the enduring legacy of the 1976 “Judgment of Paris,” the blind tasting that reshaped perceptions of wine quality and transformed the global wine industry. From statistical analysis of wine rankings to the psychology of taste perception, the conversation examines how experts evaluate wine and why even trained judges often disagree. Ashenfelter reflects on decades of wine tasting data and the role of probability, humility, and climate modeling in understanding wine quality, while Lin shares insights from her groundbreaking research on how music influences the perception of champagne. Together, they explore the complex relationship between sensory experience, human judgment, and data, revealing that wine may be as much about context, memory, and emotion as it is about chemistry and…