Can data help make food tastier and healthier?

The fundamental question about why we eat what we eat led Ganesh Bagler to explore computational gastronomy, a blend of data science and food.Bagler, professor at Indraprastha Institute of Information Technology Delhi, found that the field can transform the food landscape and be leveraged for better health and nutrition through culinary applications like recipe design, food-beverage pairing, novel recipe generation, disease-specific culinary recommendation, dietary interventions, flavour combinations and taste predictions.“I want to make food computable. But it is not just about capturing food-related data or documenting data in a digitised manner. It is about applying algorithms from computer science,” he says.In the past eight years, he has gathered data and posted it on his online open platform RecipeDB. It has 1.2 lakh recipes from 26 regions and 74 countries in a structured manner and computable fashion. Bagler and his team also have an open-source database, FlavorDB, with 25,595 flavour molecules representing tastes and odours, and DietRx, which explores the health impacts of dietary ingredients by integrating the relationship between food and key molecular agents.They are constantly updating the databases. They are also working on a technique to scan a food item via a cellphone to gauge and arrive at the calorie content in the food. medium98752941 medium98752950“Databases can tell us many things. We have answers to queries like how to get a similar flavour using another ingredient, if the original ingredient is not available locally. Similarly, one can have a list of ingredients and ask for a recipe. Or one can seek French recipes that use onion and chilli but not garlic,” says Bagler.Bagler’s use of computational technologies to find out the unique food pairing of Indian cuisine was identified as an emerging technology by MIT Technology Review.Popular belief is that recipes taste better when ingredients have a positive food pairing (similar in flavour). Bagler’s research shows that Indian cuisine is distinct because of its characteristic contrasting food pairing.One of the biggest uses of computational gastronomy is in sustainable food. “We have the carbon footprint for 600-700 ingredients. We have also computed it for recipes. At some point, governments may be interested in such data. In developed countries, customers ask about the carbon or water footprint in the recipe. Companies are building products with both in mind to serve such customers,” adds Bagler.Today, he is working with 8-10 food and beverage companies. Companies, he says, want to find out about food pairings to market their products better. Some want products that will be acceptable across the Indian subcontinent.How computing helps chefsData about our food, combined with artificial intelligence, is transforming the way scientists, chefs and individuals understand food. Paul Magelli, chief executive officer of US-based food and data analytics company WISEcode, says computational gastronomy gives us a deeper insight into what is in our food, how it impacts our health, why it tastes and smells great, and if it is sustainably produced.Experienced chefs are using technology to try different and sometimes ‘gross’ combinations to churn out delicious food.Chef Parvinder Singh Bali at the Oberoi Centre of Learning and Development in New Delhi says he had watched renowned chef Heston Blumenthal work closely with scientists to establish that chocolate and blue cheese are most compatible as their flavour molecules complement each other. “For an experienced chef, that combination may sound gross, but with science, everything is possible,” he says. medium98752905Bali wants data science to give him a historical evolution of an ingredient, its origin, varieties, traditional dishes now forgotten, and what it should not be combined with. “Certain nutritional foods lose their value when paired with other ingredients which can have a counter effect on health. I want new technologies to help me here,” says Bali.Chef entrepreneur Akshay Malhotra, who is a partner at TagTaste, an online community for food professionals to discover, network and collaborate, says they are also into digitising taste data.“I have used the FlavourDB database to better understand food pairings and Bagler’s research to know why in some cases strawberry flavour in cakes and beverages tastes like pineapple or tropical fruits. We discovered that the flavour compounds between the two are quite common. Since strawberry flavour compounds are expensive, sometimes, some companies add pineapple,” says Malhotra.