By Jackie Swift
Farmers today need to be able to access and use data if they want to carry out efficient, climate-smart agriculture. But the road to data-driven farming is not an easy one, according to Dr. Miel Hostens, Robert and Anne Everett associate professor of digital dairy management and data analytics.
Hostens, who recently joined the animal science department, has been working at the intersection of dairy farming and data science for close to two decades. He began his academic career by studying veterinary medicine in the early 2000s. At the same time, he also worked with information technology (IT).
“But I never worked with IT tools on the veterinary side,” he said. “In the veterinary world, no one used it much. Often, veterinarians would refer to IT and statistics as something beyond their skillset and would step back and say, ‘That’s for statisticians.’”
As a PhD student, Hostens realized that farms had a lot of data, but much of it was still handwritten, kept in physical files or scattered about here and there. In the best circumstance, it might be kept in a spreadsheet. “I thought, this is the twenty-first century. This can’t be true,” Hostens said. “I knew there were already all kinds of software being developed to handle data like this.”
Hostens dived in and began to build a data pipeline for raw dairy farm data — basic information such as an individual cow’s weight, age, milk production, and reproduction history. “I was trying to automatically get that data from farmers to me so I could do my research,” he said. “But very quickly other researchers wanted to get the data, too, to test their own hypotheses.”
To fulfill the need, he developed an analytical pipeline and data warehouse architecture for dairy farm data. His prototype was eventually bought by Delavel, one of the largest milking equipment manufacturers in the world, and merged into DairyDataWarehouse.com.
Since those early days, Hostens has continued to focus on creating methods and software to combine dairy farm data streams so they can be used by others for their own research. “There’s a lot of information already available in the data from farms,” said Hostens. “These farms are like continuous research tools, or research herds, that provide what’s called ‘real world evidence.’”
However, integrating the data from all those farms is a problem. “Each piece of technology has its own software, and the farmer gets locked in and can’t use anything else,” Hostens explained. “Every new piece of farm technology also comes with another piece of software. This is the situation in Europe, in the United States, all over the globe.”
Combining data from cutting-edge farm technology is central to a project Hostens is involved with called the Cornell Agricultural Testbed and Demonstration Site (CAST) for the Farm of the Future. Headed by Dr. Julio Giordano, professor of animal science, CAST seeks to develop an ecosystem of networked technologies and techniques to meet the needs of modern farms.
CAST depends on many types of data. From greenhouse gas sensors to wearable cow health monitors to GPS-guided autonomous tractors, the data streams are fast and furious. But how does an individual farmer use them in a holistic way to make decisions?
“At CAST, we are trying to come up with a methodology to combine different data streams,” Hostens said. “I believe the best way to approach this is to bring together data from, say, three different streams and then show the value of that combination to farmers and the industry. Then we can work from there to keep adding more types of data streams as time goes on.”
Hostens is also working with Joseph McFadden, assistant professor of animal science, on the Accelerating Livestock Innovations for Sustainability (ALIS) project. ALIS is a transdisciplinary collaboration that uses feed additives, ration balancing, technology and data modeling to create solutions for climate-smart animal agriculture.
For ALIS, Hostens is creating methodology for combining disparate data collected by researchers using GreenFeed units. The units measure gas emissions, such as methane and carbon dioxide, from individual animals as they eat. GreenFeed is important for the ALIS project as McFadden and his collaborators seek to track cattle methane emissions and the effects of various methane-reduction strategies.
“Many researchers all over the globe have used GreenFeed,” Hostens explained. “They receive their GreenFeed data in Excel spreadsheets from the company that owns the technology. Then they add their own data in their own format, which is not standardized. People know there’s a lot of value in those data sets. If all the data was formatted the same way, we could combine it, and we’d have the answer to a research question in days, not weeks.”
The protein needs of an increasing global population drives Hostens to combine his passion for dairy with climate change research. “All these people need protein, which they can partially get from dairy,” he said. “But we are still putting out too much greenhouse gas emissions.
“That’s why I like working on the CAST and ALIS projects,” he continued. “What my colleagues are doing is extremely important, and I can help them by speeding up that process of understanding and sharing data.”
Source : cornell.edu