Agriculture giant Land O' Lakes pressed Google Cloud into service.
The use of big data techniques to aggregate and cull information about farm production is growing, which means the need for cloud computing resources to store and analyze that data is on the upswing as well.
Case in point: One unit of the huge Land O’ Lakes farming co-op is now using Google goog Cloud Platform to design a data analytics system that will store reams of data gleaned from farmers and then allow them to analyze it.
Based in Arden Hills, Minn., Land O’ Lakes is probably best known as a big dairy supplier. It sells branded cheese, butter, and other dairy products from co-op member farms. But the company also comprises the Purina brand of animal feed and the WinField “crop inputs” division. “Crop inputs” are things like seeds, fertilizer, and pesticides or fungicides, which WinField sells to an estimated 300,000 farmers via a network of retailers.
Given Google’s push for enterprise cloud customers, it’s touting this as a big win because Land O’ Lakes overall is number 203 on the most recent Fortune 500 list with some $15 billion in annual revenue.
Get Data Sheet, Fortune’s technology newsletter.
WinField has about 200 research farms it uses to plant and test seed hybrids and crop protection formulas under different circumstances, said Teddy Bekele, vice president of information technology for the WinField division.
“If you’re a farmer in northeastern Iowa planting this year, and this year looks like the weather will be like the weather in 2011, here’s the seed hybrid that will work best for you,” Bekele told Fortune.
Much of agriculture has been driven by intuition and custom in the past, but farmers now are relying more on science and technology. “If you look at corn, which is the largest crop in the U.S., average corn production last year was 170 bushels per acre. But a farmer in Georgia got 503 bushels per acre,” he said. The co-op’s goal is to help farmers make decisions to get the best return on their land, time, and effort,
Recommendations to that end come from a lot of information sources about weather, soil, and water conditions using historical information as well as image data from satellites or planes, soil sampling, and more, which WinField has been collecting for years.
That variety of sources is both a blessing and a curse. It’s good to have a lot of sources, but the data streams and tools used to parse them were all isolated from each other. The challenge of big data, is somehow getting all that data into one accessible place.
“We needed a way to connect all that up and that’s the cloud. We needed a central place to store that data,” Bekele reflected.
So approximately 12 months ago, Land O’ Lakes started looking into the Google Cloud Platform to use for that purpose. Google’s appeal over rival clouds like Amazon amzn Web Services or Microsoft msft Azure was Google’s extensive experience with mapping.
“We have a ton of geospatial data so Google’s mapping capabilities were already there,” Bekele explained. “Any tool we worked with, whether our own or someone else’s is truly geospatial with a GIS component to it.” GIS stands for Geographic Information System, which is a computer system designed to capture, store, and parse spatial or geographic data—from maps, aerial or satellite images, what have you.
Read More: Google Plans to Boost Its Cloud Game
WinField’s needed to be able to take an image of a field and apply different overlays, one for where seeds are planted and another for where spraying is critical because seed coverage needs to extend right up to the boundary to maximize yield while spraying has to be kept away from borders or water sources.
“It’s not unusual for us to assign four to five boundaries for one field, depending on crop rotations and other factors,” Bekele noted. For that, Land O’ Lakes needed a robust mapping engine that already existed or else it would have had to build one on its own. Google was became the answer.
After selecting Google, Land O’ Lakes had a prototype application up within six weeks and has spent roughly eight months in total fine tuning and honing it with help from the Internet giant. Over the last month, Land O’ Lakes has been working with its 1,300 partners to load data into Google Cloud Storage.
Down the road, Bekele expects the system will enable farmers or their “delegated agents” to see their own data (but not that of others) and run their own analysis. He expects many will be interested in using Google’s BigQuery data warehouse and/or BigTable NoSQL database to run their own data operations.