The application of “big data” in farming makes sense in theory. More data enables better decision-making, but at the individual farm level it can be difficult to accumulate a large enough sample size for actionable agronomic analysis.
Variety XYZ performed well on one field last year, so should you grow more of that variety next year? Or was the higher yield due to nutrient availability or inputs applied in that field? How did it yield for other growers in the region? What’s the optimum nitrogen rate for this variety? Are you getting the biggest bang for your buck at your current seeding rate? These are the types of questions growers are continually seeking answers to.
Farmers Business Network — an independent California-based startup launched a year ago with backing from Google Ventures and other venture capital funds — is looking to help farmers with these decisions by aggregating agronomic data across millions of acres through a farm network.
“Our model is to take massive amounts of harvest, planting, fertility, weather and other sorts of data and put it into an anonymized dataset and use data science to identify patterns and trends that farmers can learn from,” explains Kelby Kleinsasser, vice president of business development with FBN. (continues below)
The company has two requirements for farms who become members in the network: 1) they must pay a flat $500 subscription fee annually and 2) they must contribute their agronomic data, which for most farms mainly consists of planter and yield data.
Of course, the entire concept is based on having enough acres in the system to make FBN’s analysis relevant at a local crop level. Speaking with RealAg at the Prairie Grains Conference in Grand Forks, North Dakota last month, Kleinsasser said the Red River Valley region of North Dakota and Minnesota is one of their data hotbeds, with over 500 thousand participating acres in North Dakota.
Farmers must also be able to trust the data. Kleinsasser says they have several lines of defence to maintain its integrity for situations where the neighbour’s combine yield monitor isn’t calibrated correctly or there’s some sort of anomaly.
“If the data looks suspect it’s not included. We’ll put it in your account and make it available for you but it will never make it into the master dataset,” explains Kleinsasser in the above video. “We have a data cleaning team who’s sole purpose is to clean farmer data. So if you think there’s magic that takes place at FBN, there’s not. At the end of the day, it’s a human being that’s skilled, that understands this stuff working on the data.”
FBN is not yet available in Canada, but he indicates expanding north is a natural next step. In the U.S., FBN uses government soil type maps for a base level of soil information. That data isn’t available in Canada, notes Kleinsasser.
“There are some technical challenges, but there’s really no other reason why Canada isn’t already up and running,” he says.