Knowing where our food comes from seems more urgent when there’s a bacterial outbreak like there was in March involving romaine lettuce traced to Yuma, Ariz. With the outbreak spreading to Florida and 28 other states, killing one and hospitalizing more than 60, it seems there must be a better way to track our foods’ origins.
It looks like there is.
When a 65-year-old woman falls through the cracks in our healthcare system because she doesn’t visit a primary care physician very often, she may be headed for an unplanned hospital visit. If someone studies her medical history -- if they assess the probability of hospitalization based on her medical record and numerous health conditions like congestive heart failure and substance abuse -- they could intervene.
In this increasingly data-conscious world, analytics can be used to solve problems like these and many others. That was the central message as some 376 learned at the second Florida Business Analytics Forum at the University of South Florida on Tuesday, May 15.
The university brought in thought leaders who shared how they are using analytics to solve problems. The goal was to showcase their work and the work of USF researchers to the business community.
Frank Yiannas talks about tracing the source of our food.
Among the speakers: Frank Yiannas, who oversees food safety and other public health issues for the giant retailer Walmart. Yiannas, who holds a master’s degree in Public Health from USF, says the lack of transparency is the Achille’s heel of our food system.
“It time for end-to-end traceability,” he asserts.
The solution might involve using blockchain, a digital platform Walmart employed in its own case study involving sliced mango stocked in refrigerated displays. Using paperwork, tracking it manually to Brazil took nearly a week. The time to track it using blockchain? 2.2 seconds: The time it took to scan the package. With a tool like Google Maps, even the farm could be pinpointed.
“We had food transparency at the speed of thought,” Yiannas says.
Although he initially was skeptical of it, he now believes blockchain works better than traditional methods and will bring a new era of traceability.
“It democratizes information,” he says. “Everybody that is putting data in is getting smarter together.”
Yiannas says Walmart will likely begin utilizing the technology during the next year. Perishables like produce would likely be first.
“Comingled” foods like beef from different cows and grains milled together from different sources are more challenging, he explains.
Blockchain also has built-in protection against fraud.
“If you’re going to try to cheat, you’re not a pretty smart crook,” he points out. “In the digital platform like blockchain, I basically have a digital footprint that leads back to the fraudster’s doorstep.”
In the long run, blockchain can help pinpoint where improvement can be made in the supply system, illuminating where farm producers and suppliers can hasten the flow of food from the farm to table, he says.
Solving a multitude of problems
Erich Huang talks about how data science can improve healthcare.
Speaking about healthcare was Erich Huang, Co-Director of Duke Forge, Duke University’s health data science center. He told the story of a 65-year-old woman he calls "Betty,'' who benefitted when her needs were assessed and a pharmacy tech was dispatched to assist her with substance abuse issues and her medications. Ultimately, Betty was helped and her treatment was less expensive.
Because Betty is not the only one with a problem like this, Duke built a team dedicated to data science and machine learning.
Sometimes there is bias in algorithms used in machine calculations, however.
“Machine learning is not infallible,” he points out. “In the end, there needs to be a person consuming that information and delivering that information.”
Tina Eliassi-Rad, an associate professor of Computer Science at Boston’s Northeastern University, focused on fairness in her talk.
She says computer systems respond to how they’ve been programmed, meaning they may respond better to voices similar to the ones that programmed them, or better recognize certain skin tones or face shapes.
When algorithms are used to deduce people’s characters, it could spell trouble. “You can find yourself at Gitmo [prison] because somebody’s software said you looked like a terrorist,” she says.
An effective solution involves discussing how to design a better algorithm. In the risk assessment area, scoring might be replaced by having an individual make the decision based on real-life comparisons.
Analytics and creativity
From Valliappa Lakshmanan, a tech lead at Google for Big Data land Machine Learning Professional Services on the cloud, to a three-member panel consisting of Dayana Cope, leader of Walt Disney World’s advanced analytics group; Meagan Corbett, developer of the Tableau for Teaching program; and Moez Limayem, dean of the USF Muma College of Business; they shared examples.
“The main purpose of this event,” says Balaji Padmanabhan, PhD., Director of USF’s Center for Analytics and Creativity
at the Muma College of Business, “was to bring in thought leaders in business analytics whose ideas are on the cutting edge.”
Rather than focusing on networking, the event was intended to educate.
“We want the businesses to come in and learn, so they can go out and apply,” explains Padmanabhan, a Professor of Information Systems and Decision Sciences.
Sharing this expertise is part of an effort to grow the Tampa Bay region’s business community.
“We want our region to be known for analytics -- more and more and more,” he adds. “We believe we have the right ingredients here in Florida.”
The forum showcased 17 analytics research projects at USF. Some involved stock trading, like the research of business Professor Dan Bradley and finance Assistant Professor Jared Williams studying how taxi rides from banks to the New York Federal Reserve correlate.
Or research by graduate research assistant Onkar Malgonde, business Assistant Professor He Zhang, Limayem and Padmanabhan on how to use analytics or resolve organization bottlenecks.
In partnership with the Atlanta startup, onQ
, a team including postdoctoral scholar Stephanie Andel, business Professor Gert-Jan de Vreede, Academic Director of the Master of Management Program Triparna de Vreede, PhD student Vivek Singh, psychology Professor Paul Spector and Padmanabhan is examining how adding social features to a learning program can make it more engaging.
“We want our business community to know we are doing a lot and we can be a resource for them,” Padmanabhan says.