Big data pros, cons, Connectivity will affect farmland value
JAMESTOWN -- If a farmer's equipment collects and shares data, should he be concerned that he could get in trouble if the government uses the data to find he was over-applying fertilizer or using the wrong chemical?...
JAMESTOWN -- If a farmer's equipment collects and shares data, should he be concerned that he could get in trouble if the government uses the data to find he was over-applying fertilizer or using the wrong chemical?
"I think you should, because you've proven to us that you've applied it off-label," said Terry Griffin, a cropping systems economist at Kansas State University in Manhattan, answering a question at the fifth Precision Agriculture Summit.
He added the farmer might get in trouble even if the misapplication was unintentional, done by a third party or because of an error in reporting the data. Someone could maliciously plant the false data, he said.
Griffin was one of the speakers on the "big data" concept at the fifth annual summit, held Jan. 18 and 19 at the North Dakota Farmers Union conference center in Jamestown. The event attracted about 250 people, including farmers, agronomists, manufacturers and others interested in the topic of using computers and technology to improve agricultural production and economics.
Data generated by farm machines and drones are more meaningful when shared with a third party, Griffin said. He doesn't see as much concern about people getting into an individual's system and stealing data, but "once it's in an aggregate form, in the community, yeah, we'd better have the best security possible before somebody hacks into it for sport or profit."
He said, "How many data repositories exist in agriculture today? Hundreds. How many do we need? You just need one." A single repository of agricultural data would have the most value for society, or for marketers, but that would also have limitations.
Griffin said there should be no illusions that repositories are completely safe. "If a data set has value, it will be hacked," he said.
The data has primary value for the farmer, but it's unclear when it will have "secondary value" to a third party.
Shannon Ferrell, an Oklahoma State University professor, said things are changing fast. "Our equipment can talk to itself and talk to other pieces of equipment, and talk to other people without us intervening," said Ferrell, a lawyer. "Lots of the new tractors rolling off the assembly line start talking to the manufacturer from the moment they roll off the assembly line. And they never shut up, by the way." The transmissions can be turned off, but seldom are.
True, private companies could use their "big data" potential to discover the size of a crop, and use it for a market advantage in commodities trading.
"If anyone tried to manipulate the market like that, Congress would come down on them like a ton of bricks," Ferrell said. "I really don't think there's a high probability of that happening -- at least in a visible way."
Privacy and data security are both concerns, Griffin said, and are not the same, although the two terms are used interchangeably.
Privacy means keeping others out of your data. Security means keeping from losing data. "Are you more concerned about your hard drives and losing data, or more concerned about a neighboring farmer having access to your data?" he asked.
He noted people often give information to Google Maps, but thinks the benefit outweighs the danger.
Griffin said other industries where data repositories are kept are not selling products, goods or intermediate goods. "For the most part, all of the (agricultural) data repositories today are doing just that," he said. "The group that is successful and will control the flow of data when the industry is mature may not exist yet. It may be a company we have never heard of."
What, me worry?
People might as well stop worrying about privacy because of how much information they already share voluntarily. There is primary and secondary use of data. Primary uses are for the individuals, and secondary uses rely on much more data than is currently aggregated.
"When big data are truly big, anonymity does not exist," Griffin said, flatly. "Think of it this way: If you get a big data set, and remove all of the personal level information from it, and give it to me, all I have to do is find one or maybe two more data layers, probably publicly available, combine them and then figure who (an individual in the set) is -- when data is really big. Anonymity does not exist."
He said, "In agriculture, part of the problem now is we've been stacking up the costs of participating in systems -- at least the perceived costs. But now we've done a good job of defining what all of the benefits are. The benefit-cost ratio isn't as clear as Google Maps.
"Should we do a better job of protecting our data from being lost or destroyed?"
The cost of protecting data is less than what farmers invested to collect the data in the first place. Griffin thinks the court systems will decide the value of the data.
Griffin predicted farmland value will be linked to broadband internet connectivity.
On a 10,000-acre farm, the farmer might not be on the combine, and probably is never on a tractor, Griffin said. "They have semi-skilled operators who may or may not be able to read or write, operating $200,000 machines," Griffin said. "You're using telematics to monitor those. A farm comes up for rent and you know the cell phone doesn't work, are you going to pay as much rent for it as you would a similar farm where you know your telematic technology would be beneficial?"
The definitions of broadband access are changing. In 2010, the Federal Communication Commission defined broadband access as 4 megabytes per second (MBps) of download speed -- what Netflix-watchers use. But in popular culture, the broadband upload needs have increased because of photo and video sharing.
University of Tennessee research indicates a 40-acre field with 17 pictures would need upload efficiency of almost 3 MBps. A 92-acre field with 152 pictures would need almost 5 MBps to transfer data.
An Ohio State University study has pegged transmission requirements for data on various uploads from machines -- spraying, 0.3 MBps per acre; planting, 5.5 MBps; yield data, 4.2 MBps; and soil and fertility data, 0.6 MBps.
Getting prescription data downloaded to the field is only 0.1 MBps. Significantly, the prescription files are the only ones downloaded to the field generated on computer and pushed to an applicator. The rest are uploaded.
Under the 2010 to 2014 federal definition, almost all of North Dakota, South Dakota, Montana and Minnesota had broadband (download speed of 3 to 6 MBps advertised speed) but that's a composite of all providers and individuals don't have that capacity throughout a region. Under the new 2015 definition, only small areas of south-central Minnesota have download speed of at least 25 to 50 MBps.
Meanwhile, a KSU study in 2014 showed that broadband upload speed indicates large areas of North Dakota and Minnesota have more than 85 percent coverage of 3 MBps upload coverage, while large portions of northern and northeast South Dakota and much of eastern Montana have less than 50 percent coverage.