Elena N. Naumova and Katie Stebbins of Tufts University; Dantong Wang and Fabiola Dionis of Nestlé Research, Switzerland
Partnerships between academia and industry in nutrition and health data science are exciting, but inherently complex. At their best, they bring together complementary strengths. On one side, there’s academic rigor and methodological innovation. On the other, practical application, scale, and resources. Yet, these collaborations are shaped by a fundamental tension—academia seeks to produce generalizable knowledge, while industry is driven to identify timely, action—Elena N. Naumova and Katie Stebbins of Tufts University; Dantong Wang and Fabiola Dionis of Nestlé Research, Switzerland, and consumer-relevant solutions. Rather than a flaw, this difference is a defining feature—one that requires deliberate effort to achieve meaningful, lasting impact.
Our experience building the Global Nutrition and Health Atlas with Nestlé Research taught us that successful collaborations come less by eliminating differences and more by learning how to work with them. Along the way, we encountered practical questions that arise in almost every academic–industry collaboration: What data can be shared? What should remain proprietary? How much documentation is sufficient and comprehensive? How do we balance the desire to move quickly with the academic need for verification and reproducibility? And how do we know whether what we’ve built is actually useful?
In our experience, the tension between the two camps often comes down to data. Access to rich proprietary datasets helps industry contribute meaningfully to product innovation. However, sharing data can be challenging in environments with closely guarded intellectual property. Meanwhile, academia emphasizes openness, transparency, and reproducibility. In nutrition and health especially—where evidence informs both policy and public trust—data access, ownership, and governance are especially sensitive. And new hurdles related to interpretability and reproducibility are emerging as advances in analytical methods, artificial intelligence, and synthetic data become more prescient.
But these challenges don’t have to derail a collaboration. Instead, parties should make expectations clear from the start. Rather than assuming all data should be shared, we had to work together to draw the line between what could be public and what required additional documentation or agreements. The resulting consensus helped us move forward while preserving important principles and practices for both sides.
Our experience has shown these challenges are not only manageable but also create opportunities for innovation. One clear example: the partnership between Nestlé Research and the Gerald J. and Dorothy R. Friedman School of Nutrition Science and Policy at Tufts University. This led to the Global Nutrition and Health Atlas—a publicly available platform that integrates nutrition and health data for more than 190 countries. Spanning more than 500 indicators across 26 topics throughout three decades, the Global Nutrition Health Atlas exemplifies transdisciplinary, use-inspired research. It consolidates data from international agencies, nongovernmental organizations, academic institutions, and peer-reviewed studies into six key domains—demographics, dietary intake, nutritional status, health status, health economics, and sustainable food systems. This framework allows users to navigate, analyze, and visualize information efficiently.
As a result, the Global Nutrition Health Atlas became more than a data platform—it was a shared workspace in which academic curiosity and industry experience flourished cohesively.
For our industry colleagues, platforms like the Global Nutrition Health Atlas are invaluable. Their researchers rely on robust public health data to inform the development of science-based nutritional solutions across diverse populations and life stages. The academic team, meanwhile, learned that building a useful platform required thinking beyond the scientific qualities of the underlying data. How users found, interpreted, compared, and visualized information mattered just as much.
The rapid expansion of food and nutrition information calls for new ways of data sharing and dissemination. Interactive platforms that integrate data portals and visualization dashboards can describe, monitor, and track information related to food and nutrition. Thus, building the Global Nutrition Health Atlas collaboratively demonstrated our shared capacity to translate complex data into actionable insights—an essential component of innovation in both public health and industry settings.
The project also pushed us to think differently about how we evaluate data dashboards. This led to designing a range of metrics and evaluation principles to improve data standardization and harmonization, dashboard performance and usability, information, and knowledge sharing among researchers, practitioners, and decision-makers. In other words, we learned a technically successful platform is not necessarily a successful one. Success also depends on whether people can use, trust, and find value in it.
Our work together soon expanded beyond building the platform to include broader scientific engagement and information sharing, with more than a dozen presentations at national and international venues, including the annual meetings of the American Oil Chemists’ Society. With invited talks on the potential of causal AI, the development of online surveys, and the usability of synthetic data, we addressed critical questions of reproducibility and methodological rigor. These activities reflect an ongoing joint effort to integrate academic depth with industry relevance. As we worked together on a narrative review of AI applications in lipid nutrition, we discovered new opportunities for data analytics—from nutrient interactions and lipidomics to molecular mechanisms, personalized nutrition, product development, and predictive modeling for public health at the global level.
Along the way, we also learned several practical lessons for sustaining such collaborations. Synchronizing academic calendars with industry timelines and project deadlines, for example, required careful planning—but we made it work. Establishing clear legal agreements and expectations early helped prevent misunderstandings and supported long-term trust. Perhaps most importantly, we learned that differences in perspectives were useful when they were discussed openly. Academic partners may ask whether a method is sufficiently validated, while industry partners may ask whether it’s practical, scalable, and useful. Neither question is more important. Considering both equitably can lead to a stronger solution.
The joint research also created opportunities to bring these real-world challenges into graduate education. Integrating our projects into graduate coursework allowed students to engage directly with real-world data in learning about authorship, ethical responsibilities, and the iterative nature of research from abstract submission to publication. These experiences fostered a culture of open and transparent scientific practice while giving students a glimpse into how research works in the “real world.”
We were also surprised by how effective the collaboration proved to be. We initially thought the greatest value would come from combining academic methodological expertise with industry data and resources. But we found the learning to be much more reciprocal. The academic team gained a deeper appreciation for the importance of usability, scalability, timelines, and practical decision-making. Our industry colleagues, in turn, grappled with questions of methodological rigor, reproducibility, and the broader scientific context in ways that strengthened the work. The alliance became more than a way to accomplish a project. It became a way for both teams to rethink how the project should be done.
Looking back, the strongest evidence of success didn’t come from publications and presentations. It came from better workflows, deeper understandings, stronger trust, and broader opportunities. It changed how we thought about graduate education and how we prepared students for the realities of research. Looking forward, we emphasize the need for building modular project-based courses that incorporate feedback, feedback-on-feedback, finely tuned classroom teamwork, and engaging students in the cutting-edge initiatives of the Tufts Food & Nutrition Innovation Institute.
The impact of our collaboration grew from an initial effort between teams at Tufts University and Nestlé Research into a community of more than a dozen contributors, including five generations of graduate students at the Friedman School of Nutrition Science and Policy and three rounds of participants in the Health and Nutrition Division of the annual AOCS meetings.
What advice would we give someone starting an academic–industry partnership? Begin with agreeing on the purpose of the collaboration, but don’t assume that everyone defines success in the same way. Establish expectations about data, ownership, authorship, timelines, and decision-making early. Make room for disagreements. Treat them as part of the scientific process rather than a threat to either side. Most importantly, create opportunities for both sides to learn from each other.
The strongest collaborations aren’t those where one side provides expertise and the other applies it. They are when both sides elevate their thinking as a result the co-creation.
Perhaps the most important lesson: Successful collaborations depend on a shared commitment to advancing science and serving the public good. The Global Nutrition Health Atlas platform illustrates how jointly developed resources can serve a wide range of stakeholders—researchers, policymakers, practitioners, and industry partners alike. In addressing global challenges such as nutrition, health disparities, and food systems, solutions must extend beyond institutional boundaries. Our experience reinforced that the goal is not to eliminate the differences between academia and industry, but to align them toward common purposes—advancing knowledge, informing policy, and improving health outcomes in a scientifically robust and socially responsible way.
Leave a Reply