When it comes to big data on the farm, the final destination is sunny, but the road ahead is full of bumps.
That’s according to NDSU’s David Saxowsky, a professor of agriculture who’s written on the topic of data and its coming impact on agriculture.
Saxowsky imagines a time when farmers are so well informed about their land that they can tailor seeding rates and inputs not just according to field, but according to patches within the field.
Likewise, the farmer can predict the impact of those variations, not just on yield, but on profit by taking cost data into account. But at the same time he concedes the concepts remain exactly that — unrealized concepts.
“I don’t think we’ve figured out yet how to do that,” he said.
Despite that fact, Saxowsky remains optimistic and sees big data as one of the key trends that will affect the farm of the future.
Data management is not a new conversation in agriculture. By 1994, U.S. company Satshot had already released aerial mapping software, about the same period that the first farms integrating GPS in Canada began to appear.
That said, the conversation has picked up steam as technology improves, data-based agriculture companies gain traction and practical applications emerge.
Today’s farmers are using historical and current weather to predict disease and insect risk. Farmers can identify spots with too much or too little moisture, measure the effect of input level on production and track soil nutrients over time.
Most recently, Manitoba company Farmers Edge announced that customers would be able to access daily satellite imaging, once an unheard of luxury, to measure the impact of farming practices, monitor crop stress and predict yield.
But despite this growth and the advent of the smartphone, which puts much of the necessary data in the farmer’s back pocket through a growing number of available apps, there are still major challenges in the way of the average farmer embracing big data.
Broadband internet is a basic service, according to 2016’s ruling from the CRTC, but rural internet access is still a problem in many areas.
A study out of Brandon University in late 2016 found large gaps in internet service and slow speeds along the U.S.-Canada border, through the Riding Mountain area and widespread through the northern part of agricultural Manitoba.
“A huge amount of this information, this whole logic around ‘Ag 4.0’ or digital agriculture evolution or whatever crazy exciting terms you want to put on it, it’s predicated on being able to get data on and off the farms and if a combine harvester or whatnot isn’t connected to the internet, it can’t do that,” said Evan Fraser, scientific director with the University of Guelph’s Food From Thought initiative.
The federal government announced $76.6 million in September 2016 to launch Food From Thought, tasking it with advancing the “digital revolution” in agriculture and food production. That mandate extended into big data and how it could be used to reduce inputs, monitor watersheds, isolate better-performing crops in changing climate conditions, track food through the supply chain and monitor disease threats in livestock, among other projects.
Fraser added that some companies have turned to Bluetooth, which does not need a constant signal, to link yield monitors with mobile devices, which then store the data and sync when the farmer returns to an area with internet.
Farmer’s Edge has also noticed the problem, according to its chief operating officer, Patrick Crampton. The company has introduced off-line capability to help overcome the issue.
“We’ve developed our scouting app (so) you can download those images at home into your app on your phone and then go out into the field and not worry about connectivity,” Crampton said. “GPS works anywhere. You can go to the spot that you want to scout and investigate it.”
The company faced a similar problem in Brazil, one of five companies Farmers Edge operates in, according to Crampton. There, the company adjusted its products to work off a mesh network, a system where each data node relays information to the rest of the nodes in the network. A similar system could be used in low-reception areas in North America, he said.
Different brands may also give back different data, Fraser said, raising the need for a more standardized system before true “big data,” can be used at more than just a farm level.
“A lot of the benefits that we imagine coming out of these technologies are only going to be realized if we can start pooling multiple farms’ worth of data,” he said. “The idea of identity preservation requires a whole lot of players — farmers, grain handlers, shippers, the whole value chain has to be agreeing to a common set of metrics and a common set of protocols.”
Farmers have been frustrated with inconsistent data over brands, Crampton said, adding that issue is among the company’s priorities.
At the same time, researchers are concerned about data quality should a growing number of farms begin to feed in.
Mike Duncan, a precision agriculture expert and Natural Sciences and Engineering Research Council (NSERC) research chair holder at Niagara College, is among those concerned. In a 2016 interview with the Manitoba Co-operator, he pegged the need for robust data at the farm scale.
“How often is the average yield monitor calibrated, for example?” he posed at the time. “If you want really good data, it should be calibrated before you start every field. I doubt that will happen and I bet that most yield monitors aren’t even calibrated once a season.”
Variety may also matter, according to Stephen Vajdik and Adam Gurr of Brandon’s Agritruth Research.
“Variety A may require a different calibration value than Variety B,” Vajdik said. “We have examples of variety trials where we would have come to a different conclusion simply based on whether we used the scale or yield monitor data.”
Poor data may result in poor management decisions, particularly if a trial is focused on a specific site rather than measuring the performance of one variety or another across a region, Vajdik added.
Saxowksy echoed Duncan’s concerns. Big data will require more information from more farms, but also over many years to properly map trends, he noted.
“Producers are telling me that, that they’re going to be concerned about the accuracy of their data,” he said. “But, again, we’re going to, with experience, refine those practices, improve those practices, improve the technologies. We’re just on the cutting edge.”
Farmers may face a steep learning curve as technology presents them with more and more information. As a result, packaging data has become a key role for the companies like Farmers Edge, that translate raw data into easier-to-understand insights and visuals.
Even so, the company has found the need for technicians to help customers sort through the services they’ve bought.
“Two years ago, the complaint on imagery was not enough imagery,” Crampton said. “I can guarantee you that the complaint going forward will start to be too much imagery and that’s where different approaches (fit) in terms of providing alerts to the growers, saying, ‘Hey, check out this field,’ instead of having to sift through layers and layers of information. I think any grower looking to get started needs to look for a partner that has that sort of simplification.”
Data may also hit a wall when projects designed by experts or academia fail to live up to expectations once they hit the farmer’s field, something Fraser says he hopes to avoid by integrating farmers in development.
“History is littered with examples of experts coming up with solutions to problems that farmers don’t have, or the solution doesn’t fit and whatnot,” Fraser said.
Regulation and ownership
Agriculture is one of many sectors to face the larger debate on data rights and ownership. Who owns the data? Who has rights to access it? Who has rights to share it?
The answers are still largely unclear and, at least according to Fraser, that uncertainty is putting up another barrier to farm data use.
“Many of the farmers that I’m talking to are, understandably, very suspicious about giving away the data that then becomes ‘monetarized’ and sold back to them,” he said. “I’m not necessarily saying that’s what happens, but there’s a perception that could happen, and that’s a very legitimate concern. Unless we get decent data sharing and data governance agreements sorted out, this will forever be a potential technology and never be real.”
Agritruth relates to both sides of the equation. Its name in research has grown over the last two years, driven by farmers looking for field-scale data and companies looking for third-party research. At the same time, Gurr and Vajdik are farmers.
Ownership has come up in conversation, Vajdik said, although Agritruth has avoided any major debates over unacceptable use so far.
All parties should be clear on what data is being collected, what it is to be used for and who has the right to share it when entering a business arrangement, he urged.
Answers become even more blurred when data begins to travel across international borders.
Crampton said there have been times where a copy of Farmer’s Edge data had to be held within one of the five countries they operate in.
“Working with global partners like Microsoft or Amazon, you understand how you need to store the data in certain areas and follow those regulations,” he said. “The biggest concern in my mind that we hear is, of course, around the whole data privacy issue, etc., and that’s where we pride ourselves in growers owning their data.”
He says the company aggregates data to provide insights, but growers control permissions and how data is shared. There have also been instances where a customer has requested their data remain on a Canadian server.
“If it’s their agronomist that they want to have access to their data, their local retail partner… they give those permissions and then we enable those partners to work with them,” he said.
Regulation has been slow, although Fraser says some protocols are emerging from industry and conversation is starting in government.
At the grassroots level, Ontario’s soybean growers have launched their own data management standards in a bid to access Japan’s lucrative edamame market.
At the same time, Fraser said large companies are developing in-house standardization, pointing to Monsanto’s purchase of digital agriculture company Climate Corporation in 2013.
“There’s a lot of examples of this being what the computer sciences call a very fractured landscape,” he said. “My sense is that there’s so much interest in this from major corporate players — John Deere and Monsanto and Bayer and Syngenta and whatnot — all those players working on things within the confines of their, what you might call, rose garden. It’s like Apple users. Apple has this all sorted out within Apple, but you don’t really have much interconnectivity between Apple and Google.”
Like most things around big data, it is still unclear what that will ultimately mean for farmers.
Both Saxowsky and Fraser are of the opinion that agriculture has yet to really explore the type of changes that producers are likely to see in the next five to 10 years as ownership and data rights become more clear.
“Five years ago, we weren’t talking about this,” Fraser said. “This wasn’t even a conversation. Now we’re having the conversation. Five years from now, what will this conversation look like? What will the data landscape look like? I think it’ll be very different. I think it will be more integrated than it currently is. It will be less farm specific and more generic.
“I guess the real question is will this go like Facebook or Google where there are these user agreements that we sign and our data immediately vanishes from ownership of the individual and is then presented back to us with lots of advertising and the map to help us drive our tractors.”