Industry

‘Graph Theory’ & Food & Retail Innovation

The Wall Street Journal: “Every creative cook faces the challenge of choosing ingredients that combine deliciously. This is mostly a matter of culinary experience, intuition and imagination, plus a lot of trial and error. But Big Data can help, too … Enter the food scientist and trained chef Michael Nestrud … Dr. Nestrud uses an arcane branch of mathematics called graph theory, the same sort of analysis used to pick out ‘cliques’ of Facebook friends, in which every member of the group is friends with every other member.”

“More recently, he has applied the technique to produce more harmonious combinations of snacks, main courses, side dishes and desserts in U.S. Army field rations, based on soldiers’ pairwise preferences. He also has used it to determine which snack foods should sit next to one another on convenience-store shelves, based on which items consumers tend to think of together In his current job with Ocean Spray … Dr. Nestrud searches through Twitter’s daily archives to find every tweet that mentions certain flavor-related words.”

“By seeing what else people talk about when they talk about cranberries, both during and after the holiday season, he hopes to learn more about the other flavors that consumers associate with cranberries, which may lead to novel flavor combinations for his company’s products. Data miners have an even richer treasure trove available in the form of online recipe archives. Every recipe testifies that someone, at some point, thought a particular combination of ingredients was tasty. Scientists are now using these archives to test a controversial idea about flavor pairing.”

Tim Manners/Brand X Ventures | Mar 28th, 2017 | , ,