The rise of social media platforms has remarkably influenced everyday life, including how food is represented, shared, and perceived online. Yet, limited attention has been given to how the nutritional content shared on these platforms reflects and shapes users’ dietary preferences over time. This study investigates the relationship between social media recipe-sharing and nutritional patterns by analyzing approximately 8 million Instagram posts from around 200,000 users across five languages (English, French, German, Italian, and Spanish) from 2016 to 2023. Leveraging a novel analytical framework powered by Large Language Models, we extracted ingredients from textual recipe posts and examined temporal trends in calories and macronutrients. Additionally, we evaluated recipes’ nutritional quality using the Nutri-Score classification system and assessed how user engagement varies by nutritional category. Our findings indicate stable nutritional patterns overall, with minimal shifts towards either calorie-dense or healthier options. However, significant continental variations emerged, suggesting that regional cultural factors strongly shape the circulation of dietary content online. Moreover, engagement analysis highlights geographic segmentation in food preferences by Nutri-Score categories, pointing to how cultural differences influence the popularity of foods with varying nutritional profiles. These insights underscore the value of understanding digital food discourse and engagement patterns when considering the social dynamics of dietary trends. • We analyzed 8 million Instagram posts in 5 different languages. • We employed LLMs to extract recipes from data and to compute their nutritional values. • The trends in calorie and macronutrient intake remained stable from 2016 to 2023. • Cultural factors shape online food engagement and nutritional trends. • Healthy food posts succeed in Europe, unhealthy ones in North America.
Etta et al. (Mon,) studied this question.