From Dirt To Data: Precision Agriculture At The Data-Intensive Farm Management Project

SASHA ZVENIGORODSKY

URBANA, ILLINOIS

Precision agriculture first gained traction in the 1990s, when GPS technology made it possible for farm equipment to map and manage fields with a level of detail that was not possible before. Farmers could accurately apply fertilizer or seed at different rates across a field, responding to variations in field conditions.

The promise of precision agriculture was thought to be huge, with potential to make farming more efficient and data-driven. However, there was a major limitation: farmers didn’t yet have the information needed to effectively apply different inputs variably across a field.

For David Bullock, a professor in the Department of Agricultural and Consumer Economics, part of the College of Agricultural, Consumer and Environmental Sciences at the University of Illinois Urbana-Champaign, and his brother Donald Bullock, a crop scientist at the same university, that gap sparked an idea: Instead of using precision technology only to manage fields, why not use it to generate the data needed to manage them better?

“Precision ag was everywhere in the news. The buzzwords and the hype were constant, constant, constant,” David Bullock recalls. “My brother was pretty skeptical.”

“You could hook tractors up to satellites and manage different parts of the field differently. The technology was incredible,” he explains. “But farmers didn’t know how much nitrogen to put on each part of the field. So you’ve got this incredible technology… but you really can’t use it very well.”

To tackle that problem, Donald Bullock proposed turning the field into a giant “checkerboard”. Different areas would receive different fertilizer rates, with GPS-equipped machinery applying the treatments automatically. Then, yield monitors would record the results.

“The goal wasn’t to make more money at first,” David Bullock says. “It was to generate more data.”

Collaboration Across Disciplines

The idea quickly became a collaborative effort between the two brothers.

“The reason multidisciplinary research is hard is because you’re speaking two different languages. You can’t communicate,” laughs Bullock. “But you can yell at your brother on Friday and come back on Monday and keep working, so it was a successful research partnership.”

Their initial experiments were supported by a university seed grant, which allowed them to test the approach on a real farm field. Soon after, they secured an additional grant from the U.S. Department of Agriculture to expand the project and build a digital platform that would allow farmers and researchers to run these experiments at scale.

“We weren’t trying to create an instruction manual. We were trying to build a decision tool where farmers can look at the data from their own fields and see whether it supports the ideas they already have about how to farm better.”

Launching the DIFM

With David Bullock at the helm as program leader, on-farm precision experimentation became the foundation of the Data-Intensive Farm Management project (DIFM), a University of Illinois initiative established in the mid-2010s, using large-scale on-farm trials to help farmers gather data and make better farm management decisions.

Building out the DIFM platform, however, proved more complicated than expected.

“We had a lot of ups and downs,” says Bullock. “I discovered I didn’t know anything about managing software development or databases.”

With secure funding, Bullock was able to recruit a strong team to develop the system that now powers DIFM, designed to make experimental planning easier and more accessible for farmers.

Beyond designing experiments, the system stores results in a shared database, cleans raw data, and runs automated analyses that generate reports for farmers.

“We weren’t trying to create an instruction manual,” he says. “We were trying to build a decision tool where farmers can look at the data from their own fields and see whether it supports the ideas they already have about how to farm better.”

The goal was ambitious, but successful.

“I thought, man, you could do this all over the world,” Bullock recalls. “If you make it easy enough, you can generate data everywhere.”

Expanding Internationally

After years of development, the DIFM project has expanded far beyond its humble beginnings in Illinois, now spanning more than 20 states and connecting several farmers, crop consultants, and researchers across the country. Recently, it has even spread internationally.

Building those partnerships has called for regular and dedicated outreach from Bullock.

“I talk to people a lot. I travel all over the place, talking about what we’re doing, connecting people.” Bullock explains. “That’s what I did for ten years: talking about it and talking about it. Over time you meet people who know people who know people.”

Those connections are beginning to translate into new trials abroad. “We’ve done a lot of trials in Canada,” says Bullock. “We did about 15 trials in South Africa last year.”

International expansion helps the DIFM gather diverse data, making its insights more broadly applicable and impactful for farmers worldwide.

“There are agricultural problems and difficulties all over the world,” says Bullock. “You’ve got nitrogen runoff going into rivers. There’s a huge dead zone forming in the Gulf of Mexico. Farmers don’t want to create that kind of environmental mess, and governments want to encourage better practices. The data we gather can help inform that and let us figure out solutions for these problems.”

“I’m not saying all this sensor data isn’t important. I’m saying the field trial data and the sensor data are complementary. They’re worth a lot more together than they are apart.”

The Value of Field Trials

As agricultural technology continues to expand, Bullock believes one element will always be essential: field trials.

“There’s a lot of bells and whistles out there,” he says. “There’s a lot of hype and a lot of beautiful pictures on the internet trying to get farmers to buy things.”

But without experiments that test how different management decisions affect yields, he argues, much of that data has limited value.

“You have to do field trials. If you don’t vary management and see what happens, you’re not going to learn how to manage anything,” Bullock insists. “And I don’t care how many drone pictures you take. I don’t care how many robots you run around. You have to get the data on how changing variables changes outcomes, or you’re not going anywhere.”

Bullock sees the DIFM initiative as a way to enhance current agricultural technologies.

“I’m not saying all this sensor data isn’t important,” he says. “I’m saying the field trial data and the sensor data are complementary. They’re worth a lot more together than they are apart.”

To Bullock, that combination is what makes the DIFM project so valuable within a rapidly expanding industry.

“That’s why this is a multidisciplinary effort,” he says. “That’s why I think that what we’re doing is necessary, not sufficient, but necessary to make this whole big ag data thing work.”   ∆

SASHA ZVENIGORODSKY

UNIVERSITY OF ILLINOIS