Poor diet is the leading risk factor for cardio-metabolic diseases globally, responsible for 1 in 5 deaths annually—surpassing high blood pressure and tobacco. Despite advances like GLP-1 drugs, these diseases are steadily rising with 40% of Americans projected to have pre-diabetes by 2030. The burden is even greater in low-income and minority populations, where disease rates are increasing at a faster pace. The major failure of past approaches to improving public nutrition has been the almost exclusive focus on interventions to change individual knowledge, beliefs, and habits around healthy eating. Increasing evidence shows that food choices are shaped by system-level factors outside individual control. The availability of feature-rich digital trace data from sources beyond diet and health provides new opportunities for monitoring, measuring, and modeling food systems and dietary influences, and how individuals interact with these systems. In this talk, I will present studies that leverage non-medical data together with approaches from computational social science, systems modeling, and AI/ML to gain insights into how these systems affect food choices and nutritional disparities, shedding new light on potential intervention points in food systems. Topics covered include using logistics and trade data to model foodborne contamination and characterize supply-side drivers of food access, human mobility data to examine spatial drivers of eating behaviors, and restaurant menu and transaction data to assess food quality and access. These insights inform interventions to increase equitable access to healthy diets, in collaboration with partners in policy.