Visual Cultural Biases in Food Classification.

This article investigates how visual biases influence the choices made by people and machines in the context of online food. To this end the paper investigates three research questions and shows (i) to what extent machines are able to classify images, (ii) how this compares to human performance on the same task and (iii) which factors are involved in the decision making of both humans and machines. The research reveals that algorithms significantly outperform human labellers on this task with a range of biases being present in the decision-making process. The results are important as they have a range of implications for research, such as recommender technology and crowdsourcing, as is discussed in the article.

Monitoring obesity prevalence in the United States through bookmarking activities in online food portals.

Studying the impact of food consumption on people's health is a serious matter for its implications on public policy, but it has traditionally been a slow process since it requires information gathered through expensive collection processes such as surveys, census and systematic reviews of research articles. We argue that this process could be supported and hastened using data collected via online social networks. In this work we investigate the relationships between the online traces left behind by users of a large US online food community and the prevalence of obesity in 47 states and 311 counties in the US. Using data associated with the recipes bookmarked over an 9-year period by 144,839 users of the Allrecipes.com food website residing throughout the US, several hierarchical regression models are created to (i) shed light on these relations and (ii) establish their magnitude. The results of our analysis provide strong evidence that bookmarking activities on recipes in online food communities can provide a signal allowing food and health related issues, such as obesity to be better understood and monitored. We discover that higher fat and sugar content in bookmarked recipes is associated with higher rates of obesity. The dataset is complicated, but strong temporal and geographical trends are identifiable. We show the importance of accounting for these trends in the modeling process.

Estimating the Healthiness of Internet Recipes: A Cross-sectional Study.

A government's response to increasing incidence of lifestyle-related illnesses, such as obesity, has been to encourage people to cook for themselves. The healthiness of home cooking will, nevertheless, depend on what people cook and how they cook it. In this article, one common source of cooking inspiration-Internet-sourced recipes-is investigated in depth. The energy and macronutrient content of 5,237 main meal recipes from the food website Allrecipes.com are compared with those of 100 main meal recipes from five bestselling cookery books from popular celebrity chefs and 100 ready meals from the three leading UK supermarkets. The comparison is made using nutritional guidelines published by the World Health Organization and the UK Food Standards Agency. The main conclusions drawn from our analyses are that Internet recipes sourced from Allrecipes.com are less healthy than TV chef recipes and ready meals from leading UK supermarkets. Only 6 out of 5,237 Internet recipes fully complied with the WHO recommendations. Internet recipes were more likely to meet the WHO guidelines for protein than other classes of meal (10.88 v 7% (TV), p < 0.01; 10.86 v 9% (ready), p < 0.01). However, the Internet recipes were less likely to meet the criteria for fat (14.28 v 24 (TV) v 37% (ready); p < 0.01), saturated fat (25.05 v 33 (TV) v 34% (ready); p < 0.01), and fiber (compared to ready meals 16.50 v 56%; p < 0.01). More Internet recipes met the criteria for sodium density than ready meals (19.63 v 4%; p < 0.01), but fewer than the TV chef meals (19.32 v 36%; p < 0.01). For sugar, no differences between Internet recipes and TV chef recipes were observed (81.1 v 81% (TV); p = 0.86), although Internet recipes were less likely to meet the sugar criteria than ready meals (81.1 v 83% (ready); p < 0.01). Repeating the analyses for each year of available data shows that the results are very stable over time.