Data that is accurate and accessible helps to drive innovation and progress, which was a key theme of discussion at the NNEdPro Sixth International Summit on Nutrition and Health. Data-driven policies and programmes have the potential to reorient food systems and end malnutrition by 2030, according to Andre Laperriere of Global Open Data for Agriculture and Nutrition (GODAN). The COVID-19 pandemic has exacerbated the existing food crisis, affecting production, processing, and distribution within the food system, and highlights the critical need for timely and reliable data to drive decision-making. The pandemic has affected food on the levels of production, transformation, and distribution, which presents an unprecedented opportunity for change. Using data, we can identify and learn from countries who have had the most success in reducing hunger (E.g., Armenia, Brazil, Ghana) and those which have achieved zero hunger while keeping adult overweight and obesity to a minimum (E.g., Republic of Korea, Japan). However, making practice and policy decisions involves a complicated process influenced by logic, current evidence, existing models and authorities, previous experiences, emotions, and cognitive biases, as discussed by Dr Jeffrey Bohn. Causal inference approaches could be one way to address some of these complications by merging nutrition data and scientific evidence to promote better decision-making in the context of nutrition-related communicable diseases targeted by the Nutrition Decade and the Sustainable Development Goals. Although challenges exist in all data science, there are particular challenges in applying mathematical precision in nutrition. Nutrition research considers dynamic processes that evolve and are often influenced by the process of studying them. Additionally, nutrition research occurs against the backdrop of traditional biomedical research where the randomised control trial (RCT) is considered the gold standard in proving causation. While pre-registration of data, protocol and analyses can address some of these primary challenges with research behaviour, to truly understand causation we must consider counterfactuals, which consider the context of the research (models, interventions, characteristics, and cognitive bias) for a more complete understanding. Causal inference tools can be applied to relevant, curated data to identify confounders and subsequent causal linkages. There is a necessity for the quality use of data to identify and strengthen high-impact policies and programmes for action on nutrition.