Precision Agriculture Yields Big Data Challenges

by   |   September 22, 2014 5:30 am   |   0 Comments

A straw-chomping, overalls-adorned farmer may not be the first image that comes to mind when contemplating the Internet of Things, but America’s agriculturers are some of the country’s savviest data crunchers. For years, farmers have gathered data on everything from soil fertility to annual rainfall to help maximize crop production.

Now, a new crop of high-tech tools are raising the bar, enabling farmers to crunch massive amounts of data collected through sensors to predict the best time to plant, what type of seed to use, and where to plant in order to improve yields, cut operational costs, and minimize environmental impact. John Deere’s FarmSight, Monsanto’s FieldScripts, and Pioneer’s Field360 are among the tools that allow farmers to collect planting and yield data from motorized farm equipment and input this information into a database that, when aggregated with multiple sources of anonymized data, produces detailed prescriptions.

But at a time when many industries are taking a Wild West approach to the Internet of Things, the farming world is tackling the trend’s more weighty issues, including privacy, data integration, and security breaches. Across the country’s fertile plains, farmers are closely examining how to combine seed science with data analysis without compromising their competitive edge.

Related Stories

Precision Agriculture: Deere Gets Farm Data from the Tractor to the Web.
Read the story »

Land O’ Lakes Analytics Platform Sows Sales Growth with Farmers.
Read the story »

IBM Watson Food Truck Rolls into SXSW.
Read the story »

McCain Foods’ Journey to Self-Service Analytics Includes Strong IT Role.
Read the story »

“The promise of big data is that, by putting all this data together, running the proper analytics and making use of massively parallel and super-computing technologies, we can do a better job of growing corn or soybeans,” said James Krogmeier, a farmer and Professor of Electrical and Computer Engineering at Purdue University. “But what’s holding up the promise from being fully realized is the accessibility of data – the capability of bringing all that information together into one place.”

The root of the problem, Krogmeier said, is that “farmers are a cantankerous bunch” and rely on a hodgepodge of equipment rather than risk “being locked in” with a single manufacturer. As a result, transferring data collected from one tractor to another for a comprehensive snapshot of an acre’s yields or plant population density is next to impossible.

To foster greater data integration, the Open Ag Data Alliance (OADA), a project in which Krogmeier is activity involved, is building a series of open APIs that will allow farmers to enable their hardware and software systems to communicate automatically through secure cloud services. Using this open standards software, farmers can integrate data from a multitude of sources, regardless of manufacturer.

The OADA is also developing guidelines outlining data privacy and use standards to ensure compliance with OADA principles. Even the American Farm Bureau is establishing a code of conduct on how to properly collect and crunch data on harvests. After all, farming is a highly competitive practice. With some of today’s leading Big Ag companies, like Monsanto, behind many of these data-collecting tools, how can farmers make sure their highly confidential data isn’t leaked, misused by vendors, or poorly anonymized when aggregated and repackaged for competitors?

For Iowa corn and soybean farm owner Victor Miller, the greatest risk of precision agriculture isn’t pooling his data with that of his competitors, but rather allowing a special-interest group to gain access to proprietary information.

“I’m worried about somebody that has no knowledge of my specific operation getting confidential information and then distorting what I’m actually doing,” said Miller, who uses Monsanto’s FieldScripts tool. For example, Miller says data reflecting the usage of nitrogen fertilizer, a practice heavily regulated by the government, can be easily misunderstood and misused in the wrong hands.

For this reason, Miller said that the OADA and American Farm Bureau’s efforts to establish a code of conduct around data collection and analytics is “very necessary, if nothing else, to provide us with a benchmark.”

Even farmers like Jim Broten, who rely on proprietary precision agriculture tools, face data challenges. Although Broten says that he has been able to grow more bushels of corn and grain with an equal or slightly smaller amount of costly fertilizer since launching a data analytics program four years ago, the savings and efficiencies he’s reaped are tough to quantify, making ROI calculations difficult.

“We definitely will get a payback in the long run in terms of sustainability, but it’s hard to put a dollar figure on that,” said Broten.

Krogmeier warned that it’s also possible to spoil a system’s predictive powers by filling it with too many observations or variables.

“If we build models with too many parameters, that are too complex, then we’re at risk for over-fitting and producing a model that looks very good on past data, but won’t predict worth a hoot,” he said.

Another factor that can stand in the way of accurate yield predictions: the weather. This is an area where strong “vendor” relations can really come in handy. Said Broten: “When you’re farming, you have to be fairly close to the Guy upstairs.”

Cindy Waxer, a contributing editor who covers workforce analytics and other topics for Data Informed, is a Toronto-based freelance journalist and a contributor to publications including The Economist and MIT Technology Review. She can be reached at or via Twitter: @Cwaxer.

Subscribe to Data Informed
for the latest information and news on big data and analytics for the enterprise.

Tags: , , , , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>