“The Plate,” National Geographic • Nov. 1, 2016
Researchers angling to solve America’s food waste problem are taking cues from Instagram and developing an app to measure food waste from your food pictures.
The working name for the app is FoodImage and it is based on another photo-based app that measures food consumption, called SmartIntake.
The premise, says Corby Martin, a clinical psychologist at Pennington Biomedical Research Center in Baton Rouge, is simple. SmartIntake, a diet tracker developed at Pennington, compares photos of an eater’s plate before and after a meal. The app then sends the images to dedicated servers, where trained researchers can download and analyze them for calorie and nutrient content. That photo-comparison process, says Martin, can be enhanced and put to other uses—like measuring how much food you leave on your plate and toss in the trash.
“We fully expect that SmartIntake can measure food waste, but we think it has to be modified to cover it whenever food waste can occur,” says Martin. The idea is to not only measure leftovers, but also how much food waste is generated during meal preparation—or even, whether leftovers ever get eaten. And that, says Martin, is a challenge. “It should do well with the meal waste data, but it will probably have limitations with the other data.”
Getting a close look at home food waste is important, says Brian Roe, a behavioral economist at Ohio State University, because there’s just so much of it. Each year, Americans throw out about one-quarter of the food they buy, costing an average family anywhere from $1,365 to $2,275 a year, according to a 2012 report from the Natural Resources Defense Council.
With FoodImage, says Roe, researchers would be able to see precisely how and when waste happens.
“You could say, ‘This household bought a lot of food in bulk and it was on sale—did they actually waste that?’” he says. If the answer was yes, for example, it might suggest urging customers to avoid buying in bulk.
The proposed app comes at a time when efforts to reduce food waste are getting more attention than ever in both the public and private sectors, at home and abroad.
Last year, the U.S. Department of Agriculture and the Environmental Protection Agency set a goal to reduce America’s food waste by 50 percent by 2030, while ReFED, a collaboration of organizations seeking to reduce food waste, identified 27 ways to do it. In 2013, the USDA issued recommendations for avoiding waste along the entire food chain from farm to kitchen table. And just this year, researchers in Europe released a guide to monitoring the problemwith the explicit goal of reducing it.
Still, American researchers are trying to take a cue from the Netherlands, where they’ve realized reducing food waste is a long-term challenge involving various sectors of the food system. There, a 2008 edict to reduce food waste by 20 percent by 2015 initially resulted in an increase, rather than a decrease, in wasted food. When the country measured its progress in 2011, food waste had jumped to 210 kg per person, up from 151 kg per person two years earlier.
Researchers suspected that as efficiency improved at the farm level, the supply of food increased—and lowered prices, making it more tempting to waste food. That in turn begged the question: What happens with food waste at home? The Netherlands then partnered with food retailers and ran public awareness campaigns, and saw waste drop back down, to 157 kg per person, by 2013—roughly the same rate that preceded the campaign. That’s the most recent data available.
The Pennington-Ohio State photo-app project is currently undergoing an initial feasibility study that researchers hope to finish up before the end of the year. Once funding has been secured, Martin and Roe estimate that it will take between nine and 12 months to develop a final product.
“If we can conduct studies that really tease out how the different approaches alter household behaviors, then we can help prioritize which of these will deliver the biggest impact,” says Roe. “But you need good household level data to accomplish that.”