Maria Bampasidou, Gentimis, Thanos
Maria Bampasidou and Thanos Gentimis
Every year new technologies introduced in agriculture promise to assist in day-to-day farm operations. Farmers now have access to all kinds of sensors, drones and smart apps for digital devices as well as management software that allows them to monitor the whole farming cycle. The structured data of purchasing and inventory management, seed, fertilizer, pesticide and herbicide applications, soil composition, machinery performance and financial management are coupled with aerial images, irrigation time series, weather information and more, creating a seemingly intractable tangle of data. One could say that we are in the digital agriculture era, the smart farm era or the era of big data.
There is a difference between the terms data and information. A yield monitor sensor will collect yield data in a grid-like pattern from a field and store those raw values as data points. Then, software will collect them in a time series and translate them into something useful, most of the time some sort of collection or visualization. This is what we call information. Although the raw data are valuable for informed decision making, the producer sees only a fraction of what was collected during the process. For the underlying data, it is important to know how the data were collected, who owns them and how they are used.
Most data from a field are collected by smart sensors that nowadays come as standard add-ons on farming equipment, especially from the major manufacturers. Adjustable yield monitors, variable rating and seeding equipment as well as optical sensors and soil testers are now default accessories included in tractors and harvesters. Their output is geo-located, creating various layers of data points.
Another source is images captured by drones with prices, including an appropriate camera and flight protocol software, that are no longer prohibitive. Larger farming operations also make use of satellite imagery whenever available that also includes weather information. Finally, smart sensors on irrigation and pesticide application systems can collect time- and position-tagged data in a time series. These datasets need to be “cleaned” and saved, most of the time in large databases. Gone are the days of simple spreadsheet files and personal computers because many farmers now coordinate their field operations through smartphones and tablets and store the information in the cloud.
Data ownership in digital agriculture is a hot topic. Is it the farmer who owns the data? Is it a company that owns the gadget or machinery that collects the data? Is it the entity the data is shared with? Is it the entity that analyzes the data? The American Farm Bureau Federation in 2015, together with many companies, published the Privacy and Security Principles for Farm Data. The report states:
We believe farmers own information generated on their farming operations. However, it is the responsibility of the farmer to agree upon data use and sharing with the other stakeholders with an economic interest, such as the tenant, landowner, cooperative, owner of the precision agriculture system hardware, and/or Agriculture Technology Provider (ATP) etc. The farmer contracting with the ATP is responsible for ensuring that only the data they own or have permission to use is included in the account with the ATP.
It seems natural that if the data come from the farmer’s operation, the farmer is the rightful owner. This is easy to understand if we are talking about hard copies of documents, such as financial statements and invoices. But if the information is uploaded to a consulting company’s platform, the company then has access to the data. People argue that because the data were shared, they are not exclusive anymore. (See the additional links below.)
Oftentimes, farmers unknowingly agree to share data after neglecting to read an end user license agreement when they purchase the smart machinery. Ultimately, the manufacturers of these machines end up collecting information from multiple equipment, plots and owners at every possible point in time, creating a repository of information on a grand scale with high variability and value. The data set collected this way becomes massive, and its ownership is nebulous.
On an individual farm level, current data are synthesized with historical, aggregated data and become prescriptions for the farmer, often through specialized software or consultants. But if they are collected and analyzed by an agriculture technology provider, the data can be used to improve various prediction algorithms owned by the provider. The outputs of those advanced algorithms are then sold back to the farmer with the promise of better information and better decision making. In this big-data scenario, the volume of the data is such that information on an individual plot and individual farmer will be way too small, like a pixel in a big picture, yet the picture cannot be created without that information.
People are generally comfortable in sharing information regarding production practices in a proper context. Producers share financial statements with bankers and discuss their concerns about production practices with analysts, consultants and extension agents. On the other hand, it may not be as welcomed to have their information pulled from a common repository owned by a third party because this might reveal private information about their operations and production practices. Still, in a big data world, everybody’s information will be visible to everybody else. But whether this is good or bad is yet to be determined.
Where there is knowledge, there is value, and the value of the data is not fully known. Farmers may agree to release their data if they receive some compensation. But how is that price determined? People tend to give away something when the price offered is above their reservation price. But even if that is agreed upon, economists will argue that price and value are not necessarily the same thing.
Because big data in agriculture is new relative to other industries, proper mechanisms need to be put in place to guarantee as much as possible that data are used effectively on a state and federal level, and the 2018 Farm Bill is a step in that direction. Furthermore, universities have proposed that they serve as gatekeepers of data in agriculture to avoid having a market for them. The institutions will guarantee that the data are collected, stored and transferred ethically. The LSU AgCenter is committed to this idea and is exploring partnerships with various stakeholders.
Maria Bampasidou is an assistant professor in the Department of Agricultural Economics & Agribusiness, and Thanos Gentimis is an assistant professor for research in the Department of Experimental Statistics.
(This article appears in the summer 2019 issue of Louisiana Agriculture.)
This photo was taken using a drone camera of a College of Agriculture class in digital agriculture taught by Thanos Gentimis, assistant professor in the Department of Experimental Statistics. It was taken this past spring semester (2019) by Felipe Hoffman Silva Karp, who is at the far left holding the controls for the drone. Others in the photo are left to right, front row: Karp, Prakash Dangal, Korey Nuchia, Hugh Bullard, Sumanth Vissamsetty and Benjamin Merritt. In the back row, left to right, are: Daniel Forestieri, Murillo Marins, Gentimis, Kelly Arceneaux, Loveprett Singh and Alexander Tryforos.
Dennis Burns, extension agent in northeast Louisiana, launches a UAV to collect data on plant health in a cotton field in Newellton in Tensas Parish. Burns is one of the AgCenter agents who specializes in agricultural technology and teaches farmers how to use it.