Precision Ag & Big Data Learning

Precision Ag & Big Data Learning

There is ceaseless talk today about “big” data in agriculture but rarely about the learning that must accompany it. Big data refers to the voluminous flow of data that is being generated by machines, devices, and software programs. It is a product of the ever expanding computer and information age. The continued addition of new technologies, such as drones, in precision agriculture contributes to a growing body of data that is being collected and disseminated at a higher frequency and to a broader audience. Recipients of big data in many cases do not know what to do with it and therefore may need to seek “big” learning. Big learning is an outcome of the second information revolution and is best understood in a historical context.


The first information revolution started with the invention of the movable type printing press in the mid-15th century. Some historians refer to the creation and circulation of printed books at the end of the medieval period as “new” learning. In a matter of years, the transfer of knowledge went from handwritten documents and oral history to the printed page. According to Wikipedia, over eight million books had been printed by the turn of the 16th century. By the end of the 16th century, an increasing fraction of first-printed books were about botany and agriculture. Experience with growing all kinds of crops in different environments was systematically catalogued for the first time. A progressive farmer at that time had only to learn to read in order to gain access to the wondrous worlds of animal and crop “husbandrie.”

In a Food and Agriculture (FAO) publication entitled “Improving Agricultural Extension: A Reference Manual,” there is a chapter by Gwyn Jones and Chris Garforth on the history of agricultural extension. The authors note that the first practical attempt at “university extension” started a little over 150 years ago in England. University educators began to provide lectures on a range of subjects important to a rapidly industrializing urban population. Agricultural extension came into being at the end of the 19th century in England and was started about a decade later in the U.S.

The Land Grant Impact

After the passage of the Morrill Act in 1862, land grant institutions were established in each American state to promote the education of farmers. Colleges given federal land grant funds began setting up experimental plots and inviting farmers to review the results. In quick succession during the first two decades of the 20th century, several states created county extension agents and forerunners of what would become known as 4H clubs. Educators at that time understood the importance of education with each introduction of new technology or practice in agriculture.

Aided by federal support for land grant colleges throughout the U.S., the extension service became part of nearly all agricultural departments. By the end of the first quarter of the 20th century, it was not uncommon for new assistant professors to have all or part of their appointment dedicated to extension duties. Throughout most of the first half of the 20th century, crop yield trials, conservation studies, and the evaluation of new equipment became commonplace. At the end of this period, chemical applications emerged as the major choice for pest control on most crops. At the same time, agricultural companies began funding their own and university research, further expanding learning in agriculture.

The use of radio and television for transmitting agricultural information became widespread by the 1950s. Most of the programming on these media was during the early morning hours so that farmers could receive timely information before going to the field. By the second half of the 20th century, extension professors became extension specialists with a limited focus on one or two crops but a broader range of expertise. They became adept at new technologies, such as modeling on mainframe computers to support production decision making, and setting up sophisticated environmental monitoring equipment in crop canopies. New learning, such as integrated pest management (IPM), was implemented to improve the timing of chemical sprays and reduce their number in order to save farmers money and limit their environmental impact.

The advent of the desktop computer, communication modems, and cell phones in the last three decades of the 20th century personalized the technology for both private and extension personnel. These computer and communication technologies coupled with the Internet in the 1990s initiated a second information revolution. The printed page slowly gave way to the electronic page and the circulation of information in society reached levels unachievable with hard copy. As access to the Internet became faster and cheaper, computers became mobile, and cell phones became smart, a deluge of data and other information from many sources became available for everyone in society. This deluge became known as “big” data.

Dealing With ‘Big’ Data

With big data becoming a reality, the next challenge was doing something with it. It became clear that big data would require “big” learning. That is, like learning to read, individuals would have to learn new skills on mining data to reveal the knowledge embedded in it. Armed with this knowledge, individuals would have the opportunity to do things in ways that are innovative and efficient. In the case of precision agriculture, the mining of big data could support more precise variable-rate applications both in time and space. Similarly, a systematic investigation of big data may allow for better production plans by being aware of the synchrony of crop and pest stages during certain periods of a growing season. Lastly, big data may provide evidence to support agricultural practices arrived at intuitively.

In the coming years, big learning will have a number of old and new tools to assist in the perusal of big data. Artificial intelligence, neural networks, simulation training, pattern recognition software, and deep learning algorithms are just a few examples that come to mind. Furthermore, big learning may eventually be assisted by intelligent machines in the field that do real-time analysis as a necessary precursor for identifying more favorable management choices.

PrecisionAg magazine and focus on new technologies for precision agriculture. Some of these technologies will require learning. In the case of new technologies that generate big data, big learning may be necessary.