Big data in Canadian agriculture? Not yet

Precision agriculture proponents are convinced that the technology will pay off, but that data collection needs to be standardized and not require farmers’ time during the busy seasons

Watching over the fencerow to see how the neighbour is growing that nice crop is a long tradition in farming. Are we heading into an era where you can look over the digital fencerow to look at the neighbour’s data on nutrient placement, seeding rates, variety selection and combine speed?

Increasingly each piece of equipment in the field has data-gathering capabilities. Combined with ever-cheaper computing power, there’s the prospect for “big data” to make big strides in farm productivity.

It’s a strategy that’s worked in a lot of other industries. For example, manufacturers use data collected about the vibrations of their packaging machines to predict when they will fail, thus avoiding lengthy downtime. Others use data to reveal hidden associations that can equal market opportunities, or great efficiencies that can lower cost and increase profitability.

But getting to big data is, first and foremost going to require a commitment to small data — the data collected at the farm level. Unless growers gather it and manage it correctly, we’ll never get there, says one Ontario grower who’s been working with the technology for years.

Rick Willemse, who farms near Parkhill, says he’s always understood his data would have value to his farm, even if he couldn’t necessarily get the full benefit of it now. Initially he used it to set up production zones and capture yield data to prove one way or another whether those efforts were paying off.

“It’s important because I need to know what’s happening in my fields,” he told Country Guide. “But it hasn’t always been easy, especially not in the early days — I spent a lot of time sitting in the headlands of my fields setting the jobs up.”

Each field operation is a “job,” and each gathers data. When Willemse is sowing the crop, for example, he’s gathering data that includes the as-applied information on all his spring inputs. Then when he does in-field operations he gathers the same sort of information for weed control and fungicides.

At harvest time, his yield monitor tracks how yield varies throughout his field. At the end of the season he’s left with a complete picture of every operation in each of his fields, and the ability to see what worked and what didn’t — in raw data form.

He keeps this information because it gives him that clear picture — but he’s hoping eventually to take it one step further and really up his crop-management game. The question remains, however — when will the reality catch up to the potential?

Not yet

One almost gets the impression Mike Duncan is tired of answering the questions. What’s “big data” mean for agriculture? And when will it really take off?

Not because he’s uninterested in the topic — Duncan is a Natural Sciences and Engineering Research Council (NSERC) research chair holder at Ontario’s Niagara College who has spent the past 10 years working on precision agriculture. Rather, he’s a bit weary of having to explain how far the business still has to come, beginning with even the most basic of concepts like what big data is, and accepting that agriculture isn’t even close yet.

“We’re not really truly talking about big data; not like most people talk about big data,” Duncan explained. “If you want to see big data, go to the Interac banking centre. It’s a bunker in London (Ontario) and you can sit there and watch millions of transactions flow in every hour. That’s big data, and we don’t have anything even close to that.”

The problem, Duncan explains, is that in the case of Interac and other big business endeavours, there’s a significant amount of data discipline. Standards are set for calibration, collection, management and analysis. As a result, reasonably clearcut conclusions can be drawn. In agriculture you’re dealing with a lot of independent operators with varying views on the importance of data collection and varying methods of collecting the information.

“How often is the average yield monitor calibrated, for example?” Duncan said. “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.”

Then there’s the question of cleaning up the inevitable statistical noise that’s going to be present in the reams of data that will be collected. For example, data cells sometimes just don’t add up — a cell that says corn yielded a bushel to the acre might be right beside one that says the same field just yielded 2,000 bushels an acre.

“If you want to be able to really use this data, and have it be meaningful, you’ve got to clean it up and remove those outliers,” Duncan said.

That’s going to take a lot of time and effort, however, and so far nobody seems to be doing it. Certainly the farmers who are collecting it aren’t likely to do this, because they’ve already got a lot on their plate and most find even capturing the data to be time consuming enough that most of them don’t do it, Duncan said.

“I can understand why they’re making that decision,” Duncan said. “When you think about it, why should they? What we’re telling them is, ‘go to the time, trouble and expense of collecting all this data, and we promise, eventually it will be worth something.’”

Big potential

Wally Orr thinks data management will be key to successfully running a really big operation.

“I can see a time when this data forms part of the value of your land when you go to sell it,” Orr told Guide during a short break from fixing a combine on his Pilot Mound, Man.-area operation.

He said his affinity for gathering the data initially began with the desire to show “proof of concept” for the rest of his precision agriculture program, including his variable-rate applications. Doing so helped prove to his own satisfaction, that the time and trouble were more than worth it.

“I am 100 per cent sold on variable rate,” Orr said. “In fact, if the satellites go down while I’m seeding, I’ll sit in the field until they’re up again. It no longer makes any sense to me at all to seed a crop without that data and prescription.”

There’s not quite such a direct benefit to other applications just yet, but Orr says he’s staying the data-collection course for a couple of reasons. First and most importantly, he thinks the data will eventually have just as much value as the variable-rate prescription data he’s using now, and his experience to date suggests that could add up to something pretty significant. And secondly, he’s enjoying the process.

“I downsized my farm a bit recently,” Orr explained. “I joke that I have a small farm run by an old farmer. Things like this keep your mind active and it’s a very interesting concept.”

Orr’s precision agriculture consultant says he’s on the right track and, like any major transition, it’s going to take some time and there will be a learning curve involved. Wade Barnes, one of the founding partners in precision agriculture consultancy Farmers Edge, got his start as an agronomist in Pilot Mound and has worked a lot with Orr over the years.

Barnes says he can understand why farmers are concerned about data integrity and wanting to make sure they’re getting fair value for sharing their data — but ultimately he’s hoping they’ll come to realize it’s in their own best interest to move in this direction.

“I am absolutely convinced this is the next big thing in crop production,” Barnes said. “It will have an impact that’s just as profound as the Green Revolution and the GMO revolution.”

Why share?

Barnes says the value of the looming revolution will only be realized if the system to collect the data is seamless and effortless for growers. The last thing anyone wants to worry about in the middle of a busy season is swapping USB sticks and uploading data to a PC in the farm office.

“You really need it to be automatic and self-maintaining for the grower,” Barnes said. “You need to be able to store that data until the machine hits the yard, for example, and sniffs out a WiFi connection, and then automatically uploads it to the Cloud where it can be stored and processed.”

Barnes says that in an ideal world growers could forget about the data until they had the time and inclination to delve into it during the off season.

A good service provider — and Barnes admits he hopes his company will become the preferred supplier to growers — would then have the opportunity to show them what the system could do, including data analysis, benchmarking with similar farms and perhaps even management recommendations.

He says only by amalgamating their data will farmers be able to truly harness the potential benefits.

“What I tell farmers is, ‘Don’t hoard your data. Share it,’” Barnes said. “By sharing it you’re not going to be giving yourself a smaller slice of the pie. You’re going to help increase the overall size of the pie and you’ll all make more money.”

He says getting a better handle on what’s actually working and not working will give growers a much better basis for decision-making.

“Do you really want to be making decisions that cost you this much just on your gut instinct?” Barnes said. “A lot of farmers do really well at that, but I think if you give those same growers better information, they’ll make even better decisions.”

About the author


Stories from our other publications