How Tyson’s CTO Is Using Big Data to Forecast COVID Spread

By Lilly Milman |  February 25, 2021

The executive leadership team at Tyson Foods always tries to work on SMART goals — and that goes beyond intelligence. These goals are specific, measurable, attainable, reportable, and trackable, says Scott Spradley, Executive Vice President and Chief Technology Officer at the company. And every year, new goals are made with the intention of improving upon the previous year’s efforts.

Scott Spradley, Tyson Foods

But in 2020, there was just one, big goal: ensuring safety among employees during the COVID-19 pandemic. 

Because food distribution was labeled an essential business, Tyson Foods only shut down its plants to undergo deep cleaning and to install safety measures. After that, it was back to business. Right at the start of the pandemic, Spradley and his team began analyzing data to predict how infection might spread within production facilities.

In a recent interview with InnoLead, Spradley shared how he helped the organization make the switch to an agile mindset, how his team built a data model to predict infections, and how Tyson is approaching automation. 

Switching From a Waterfall to an Agile Mindset

When Spradley arrived at Tyson in 2017, ending his nearly nine-year tenure at Hewlett Packard Enterprise, he was used to moving at speed, he says. But the infrastructure was not yet in place at Tyson to move as quickly as he wanted to. 

“There are certain things that you need to put in place to start driving change because you’re conditioning your workforce — both your direct workforce, the labor that is going to drive some of the transformation, as well as the workforce that’s going to benefit from the transformation,” he says. “You really want to try to condition them [to get used to the idea] that we’re going to do things significantly faster.”

That’s why “immediately” switching from a waterfall mindset, which focuses on a linear and sequential development process, to an agile mindset was one of the first things he did on the job. 

While he had been working on changing mindsets in the company for years, Spradley says the pandemic showed the rest of the executive leadership team how important a full-scale digital transformation was. The situation gave him and the data scientists on his team a chance to show how much quicker the company could react to the spread of COVID with the help of cloud technology, machine learning, and an agile mindset. 

“COVID gave us a chance to show Tyson [what our team could do], because there were those people who were still holding on to their Excel spreadsheets. As I jokingly say, ‘You were gonna pry it from their cold dead hands, because that’s what they know,'” Spradley says. “We [the technology team] can be agile, but if the company’s not going to be agile, that’s not good for us. And that’s not good for the company. We need to get the whole company thinking in an agile manner, so that we can move faster.”

Using Data to Predict COVID-19 Spread

Typically when one thinks about employee safety, a company’s technology group isn’t the first to come to mind, Spradley says. They think of a health and safety group, or of plant managers — but it was the Tyson technology team that was on the case by January 2020. 

The team created a dashboard using a Richard’s curve model, a widely used growth model. The dashboard updates information in real time using data ingestion automation — a process in which data from multiple sources is automatically siphoned into one place. The dashboard takes in information including employees’ ages, their county of residence, the amount of COVID-19 cases in those locations, and more. Then, it creates a forecast of what plants are expected to become hotspots. 

This helped the company speed up the response to infections, including implementing rules around wearing PPE to slow the spread and hiring employees from areas with less infection. According to Spradley, Tyson could now act preemptively, instead of waiting until symptomatic employees receive positive COVID-19 test results to react. 

Reducing Risk of Injury and Improving Quality with Automation

In addition to improving safety conditions in relation to COVID-19, Spradley’s team is also using automation technology to reduce injuries and improve quality. For example, he says, there are certain jobs at a factory or plant that have higher risk for injuries. “So, if you can mechanically automate that…that’s a good thing,” he says.

The company has also begun implementing computer vision in plants where multiple products are packaged on the same factory line. Traditionally, employees had to identify the different products and check them for quality. Computer vision completes the process faster and with more efficiency.

Spradley hopes to return focus back to some of these areas after COVID-19. 

“To be very candid, these jobs that we have in these plants, there’s high turnover. The employees are very smart people, and they recognize ‘that’s a high risk job,’ or ‘that’s a low risk job.’ So, if you can put automation in place…to reduce the risk of an injury, then an employee is not going to turn over and they feel safer. That’s better for us, and that’s better for our customers…because you get more consistency in the product.”