Figure 1
From: Network analysis and data mining in food science: the emergence of computational gastronomy
Flavour network . Culinary ingredients (circles) and their chemical relationship are illustrated [4]. The colour of each ingredient represents the food category that the ingredient belongs to, and the size of an ingredient is proportional to the usage frequency (collected from online recipe databases: epicurious.com, allrecipes.com, menupan.com). Two culinary ingredients are connected if they share many flavour compounds. We extracted the list of flavor compounds in each ingredient from them and then applied a backbone extraction method by Serrano et al. [6] to pick statistically significant links between ingredients. The thickness of an edge represents the number of shared flavour compounds. To reduce clutter, edges are bundled based on the algorithm by Danny Holten (http://www.win.tue.nl/~dholten/). © Yong-Yeol Ahn, Sebastian E Ahnert, James P. Bagrow, and Albert-László Barabási.