WESTERN PRODUCER — New artificial intelligence programs provide advanced analytics that can help agricultural companies improve their products.
But the insights these programs provide can only be as accurate as the information they are trained on, so there’s a need for ag companies to improve the management of their data.
Microsoft’s Azure FarmBeats has worked for years on the aggregation of agriculture data sets to help build AI and machine learning (ML) models.
FarmBeats is being retired, but some insights and methods it created will live on in Microsoft’s Azure Data Manager for Agriculture.
Claudia Roessler, program manager for agriculture at Microsoft, said Azure Data Manager connects farm data from disparate sources to help companies use the information to develop and improve their digital ag products.
“We’re working with the companies that have the agronomic expertise and are building out data models and data science in agriculture, and can provide those services to growers,” Roessler said.
Previously, companies often relied on in-house expertise to manage its data, but it can be complicated to run analytics on data that are often sitting in different data silos that do not easily connect to each other.
“The different data silos could be weather data, satellite data, data that’s streaming of equipment or sensors in the farm. We want to help them to accelerate the development of solutions by providing an analytics ready data set. So, it’s a clean data set that’s been transformed and that can be used to run machine-learning models, for example, on it.”
She said most machine-learning models require one or more data inputs to run predictive analytics.
“And then organizing this data in a two spatial and temporal manner. So those models run, for example, over different types of equipment that’s running on farms. That’s really the work that is something we help with and that’s what we’re providing to those organizations.”
These kinds of data models work best when run over multiple farms because patterns and insights only emerge with large data sets.
This is why Microsoft works with large companies like input providers, OEM equipment providers that provide prescriptions to their tractors, or with farm co-ops.
An early adopter of Microsoft’s cloud technology is Land O’Lakes, which relies on the technology for its precision ag and Truterra service.
Truterra provides insight into how different agricultural practices impact water, nitrogen and carbon on a farm, and it enables farmers to track soil carbon sequestration and participate in carbon markets.
During the launch of Azure Data Manager for Agriculture in March, Bayer announced a program called AgPowered Services, which is licensed on Azure’s Data Manager.
Brad Roberts, marketing manager of cloud solutions at Bayer, said AgPowered Services is a suite of ready-made digital products that are built with data that’s within Azure Data Manager.
“On the foundational level, we have the Azure Data Manager for Agriculture cloud that’s being partnered with, and this hosts large data pipelines, the raw data. You can imagine large sets of weather data, satellite imagery, soil information, equipment details,” Roberts said.
“AgPowered Services are a category of products that take that base data within the cloud, and builds something tangible that can be utilized and purchased and put to work for a company.”
One of the AgPowered Services products Bayer released is crop-water-use maps that use data from weather and satellite imagery to create metrics behind how much water a crop has used or lost within a 24-hour period.
This can help growers start a better irrigation prescription.
However, Roberts said Bayer is focused on a working with other business with the cloud service.
“It’s building this infrastructure and this ecosystem to where organizations can then tap into this data. A buzzword right now is interoperable data, but it’s data you can mix and match and build something. That is very much our vision that other organizations are going to build on this to provide additional value sets to farmers.”
Bayer will use this same approach for its Climate FieldView product, which will end up a customer of AgPowered Services.
Roberts said there’s a lot of infrastructure and engineering maintenance that happens with a lot of large data pipelines.
“With a product like FieldView, a lot of time and energy is spent making sure the data is flowing correctly,” Roberts said.
He said having Microsoft work on some of these pipelines, such as weather and satellite imagery, frees up FieldView to focus on providing insights.
He said AgPowered Services will help business interested in the implementation and tracking of sustainability programs with increased traceability and supply chain transparency.
It will be easier for these companies to document what’s taken place in a field to help with sustainability reporting, as well as enable better in-season decision making.
“So, if something isn’t meeting sustainability measures they have the ability, now because they know what’s happening in season, to make different decisions so that they’re adhering to those sustainability goals,” Roberts said.