Can Big Data Help Build a Better Food Bank?
It's a simple enough idea: People who have enough to eat donate canned goods and other nonperishable items to the food bank, and their hungry neighbors, some 50 million Americans, are able to bridge the gap between pay and food stamps with the donations. Except, when charities are running food distribution programs across multiple municipalities, serving thousands or even hundreds of thousands of people, it's not quite as simple as one person putting a can on the shelf and another taking it off.
Engineering professors and students from North Carolina State and North Carolina A&T State University believed they could improve what can be, in practice, a Byzantine food bank network—and enable a greater number of people to be fed with fewer donations—by using big data analysis and streamlined systems of distribution. They tested their hypotheses by partnering with one of the region’s largest food banks, Raleigh's Food Bank of Central & Eastern North Carolina, and working for the past couple of years to overhaul its system.
“Over the past seven or eight years, our distribution has gone from 30 million pounds to nearly 54 million pounds at the end of fiscal year 2013,” says Charlie Hale, vice president of operations and IT for FBCENC, which serves 34 counties. “I have got to believe some of that has to be attributed to these engineers who have helped us get more turnover in our space.”
By turnover, Hale is referring to the ability for the food bank to quickly receive donated food into warehouse space and quickly move it out to the partner agencies that need it. The FBCENC is able to turn over the food stock in its 115,000 square feet of storage every nine days or so—roughly 29 times a year.
The positive results, food bank officials say, has resulted in a renewed focus on strategies that are based on sound data and solutions that are helping the FBCENC waste less food and serve more people.
How did they facilitate so much improvement in distribution? Julie Ivy, an industrial and systems engineer at North Carolina State University, and Lauren Davis of North Carolina A&T State University put together a team to collect food bank data, analyze it, and create computational models of supply and distribution processes. What the team found was that service agencies in some counties may have a higher need for food donations, but they experience difficulty accessing their “fair share” of food from the food bank. The team was able to identify these “bottleneck counties” and come up with solutions to increase their capacity to claim the food they need—such as satellite and mobile food distribution centers. In the meantime, any food that cannot be claimed by a particular agency can be distributed elsewhere instead of thrown out, thereby reducing waste.
“At the end of the day, that just means there’s more food available for those who are in need than ever before,” Hale says.
The systems improvement, engineering troubleshooting, and data crunching that improved operations at the FBCENC may be able to help other food banks across the country increase their capacity to provide emergency food assistance to those in need. And these institutions could certainly use the help. A report released last week by the United States Conference of Mayors reveals that despite an improving domestic economy, the burden on emergency food providers is still high in many places. Seventy-one percent of cities surveyed in the report said emergency food requests had increased over the last year, with a 9 percent increase in food poundage distributed, on average.
Hale says he’s eager to continue using the processes and data introduced over the last few years, and has even sought out engineering help for other problems. Heartened by the assistance he’d received from Ivy, Davis and their classes, last winter Hale emailed the head of the mechanical engineering department at NC State, hoping for some input on another problem: The food bank was throwing out a huge number of dented canned goods from a particular donor because it was unknown whether or not they were leaking. The engineering students got to work, and designed a vacuum chamber that identifies whether a dented can is, in fact, punctured.
“We’ve been able to make that donor’s donation go further by saving more of their product,” Hale says.