Big data, small returns

Farmers provide oceans of data but get back few insights they can implement

Keena Lykins
6 December 2017

3 min read

Faced with the daunting prospect of 2 billion additional people to feed within the next 30 years, agricultural companies are scrambling to produce more food per acre.Blanketing fields with sensors that tell farmers exactly what every field needs every minute of the day could achieve that goal. But the gap between vision and reality is large.

By 2050, the United Nations projects that global population will reach 9.7 billion – 2 billion more people than are alive today. One potential solution for feeding them: ongoing, real-time collection of field-specific data on seeds, soils, fertilizers, pest pressures and weather to unlock bigger crop yields per acre.

Dozens of players, from equipment and supply companies to crop and soil specialists, have stepped up to capture mountains of farming data. So far, however, because they lack the sophisticated processing and analysis needed to extract specific recommendations from the raw numbers, farmers are realizing little benefit.

“People are collecting data; equipment manufacturers are collecting huge amounts of data and they are using it,” said Graham Mullier, the Reading, UK-based head of Data Sciences, R&D Information Systems for Syngenta, a Swiss-based agrochemical and seed company. “The question is, ‘How much of it is of value to the growers?’”


“The lack of actionable steps is why the agriculture industry is lagging behind in adopting big data,” said Matt Rushing, vice president, Global Crop Care for AGCO, a US-based agriculture equipment manufacturer whose brands include Challenger and Massey Ferguson. “Farmers like to farm. They don’t want to dig through spreadsheets or reams of data looking for insights to improve their operation. They want to see this information offered as actionable recommendations by their service providers.”

A first step, Mullier said, is to open the collected data to everyone so that it can be compiled and analyzed. “Open data movements and organizations, such as GODAN (Global Open Data for Agriculture and Nutrition), are good examples of people trying to get together and solve a problem,” he said.



But making the data widely available isn’t enough.

“We’re getting to a place where opening the data without some way to categorize or qualitatively describe the data doesn’t make a lot of sense,” said Rich Wolski, professor of Computer Science at the University of California, Santa Barbara, and co-director of SmartFarm, a research project investigating how to design and implement an open source, hybrid-cloud approach to agriculture analytics.

“The growers that we encounter are very, very focused on their problems,” Wolski said. “They don’t want to spend a lot of time thinking about data analysis. You need to throw up a map. It sounds easy, but it’s hard to distill statistical models down to a red light/green light.”


The potential of big data to change farming practices can be seen in the impact of precision agriculture (PA), which has been widely adopted because it offers a tangible benefit, AGCO’s Rushing said.

Like digitalized agriculture, PA employs information technology, GPS, sensors, soil-sampling, software and telematics to identify the best combinations of seed, water, fertilizer and agrichemicals to achieve the ‘Four Rs’—the right inputs, in the right amount, at the right time, in the right place. Unlike digitalized agriculture, however, PA recommends action steps to a specific farmer based on conditions at the time of measurement. Because PA captures data at a specific moment in time, conditions can change between snapshots.

Soil analysis consultant Jim Yager gathers electroconductivity measurements in a test orchard at the University of California, Santa Barbara, Sedgwick Reserve for use in developing a hybrid-cloud approach to digital agriculture analytics. (Image © SmartFarm)

In a current SmartFarm project, however, Wolski’s researchers permanently installed sensors throughout a citrus grove to monitor conditions around the clock. When coupled with historical weather data, they hope to provide advanced warning of where a grove will experience frost, enabling growers to concentrate frost prevention water, fans and heaters in the most at-risk areas rather than the entire grove.

“Frost prevention is very expensive, so the more focused we can be on where the frost will appear, the more money we’ll save,” Wolski said.


Outside of experimental fields, however, growers cannot yet harness always-on, real-time data-collection technology.

“The technology isn’t quite there yet, but we are looking at what the analytical models would be,” Wolski said. “One of the things we found out while working with folks in North Dakota is that many farming problems are regional or very farm-specific. You need data from California to solve California problems, and you need data from North Dakota to solve North Dakota problems.”

Agricultural technology entrepreneur David Baeza instructs University of California, Santa Barbara students and staff on fine points of drone flights to collect aerial imagery for use in precision agriculture. (Image © SmartFarm)

Bob Avant, program director of AgriLife Research at Texas A&M University in College Station, Texas, said many companies are looking at different ways to apply data technology to farming, but farmers are too busy tending their crops to have time to invest in growing and harvesting data, too.

“We’re really talking about artificial intelligence that can look at data trends and come up with conclusions,” Avant said. “Data is the root of the future of agriculture, but we’re not there yet.”

Related resources


Register here to receive a monthly update on our newest content.