219
perfect data generally is expensive to acquire and, given big data approaches,
often is not necessary to produce information that can improve decision making.
• For example, it might be technically possible to put sensors to measure actual
traffic along every mile of every road in the country. However, the cost of the
sensors and the underlying system to aggregate and communicate that information has been prohibitively expensive. However, the use of proxy information
(primarily from cell phones and sensors put in place for other purposes) allows
for useful predictions of real-time traffic conditions to be created and communicated at low cost.
• Although there are numerous mathematical and statistical approaches available
to data scientists, a key principal is the use of Bayesian inference and conditional
probabilities (Polson and Scott 2018). Essentially, through the analysis of very
large amounts of relevant data, analysists can predict with confidence that the
presence of certain factors indicates that the condition of interest exists. As a
simple example, if it is 8 am on a weekday morning (that is not a holiday) and
cell phone signals along a major highway are moving very slowly from one
tower to the next, it is likely that there is heavy traffic along that highway.
• Of course, such a prediction is probabilistic and may not be accurate in each
circumstance, especially if there is a change in the underlying conditions.
However, with careful analysis and implementation, data that is not perfect can
be effectively employed to improve decision making in agriculture. For example, consider the large maize farmer who receives satellite maps of the fields for
which the farmer is responsible. Colors are used to identify conditions in the
field, with green indicating heavy vegetative growth. As one farmer reported to
the chapter’s author, an area marked in heavy green means either that the crop is
doing really well or that there is a heavy infestation of weeds. In either case, it is
worth the farmer’s time to physically investigate.
• In agriculture, as in most fields, descriptive efforts have been most common and
even those are relatively infrequent. However, within production agriculture,
knowing what has occurred – even if very accurately and precisely – may not
provide useful insights as to what should be done in the future.
• Production agriculture is complex, where biology, weather, and human actions
interact. Science-based methods have been employed to discern why crop and
livestock production occurs in the manner in which they do. Indeed, relative to
the big data topic, it might be useful to consider this as the “small data” process.
• The process starts with lab research employing the scientific method as a systematic process to gain knowledge through experimentation. Indeed, the scientific method is designed to ensure that the results of an experimental study did
not occur just by chance (Herren 2014). However, results left in the lab do not
lead to innovation and progress in the farm field. In the United States, the USDA,
Land Grant universities, and the private sector have collaborated to exploit
scientific advances. A highly effective, but distributed, system emerged where
knowledge gained in the laboratory was tested and refined on experimental plots
and then extended to agricultural producers.
8 Digital Technologies, Big Data, and Agricultural Innovation
perfect data generally is expensive to acquire and, given big data approaches,
often is not necessary to produce information that can improve decision making.
• For example, it might be technically possible to put sensors to measure actual
traffic along every mile of every road in the country. However, the cost of the
sensors and the underlying system to aggregate and communicate that information has been prohibitively expensive. However, the use of proxy information
(primarily from cell phones and sensors put in place for other purposes) allows
for useful predictions of real-time traffic conditions to be created and communicated at low cost.
• Although there are numerous mathematical and statistical approaches available
to data scientists, a key principal is the use of Bayesian inference and conditional
probabilities (Polson and Scott 2018). Essentially, through the analysis of very
large amounts of relevant data, analysists can predict with confidence that the
presence of certain factors indicates that the condition of interest exists. As a
simple example, if it is 8 am on a weekday morning (that is not a holiday) and
cell phone signals along a major highway are moving very slowly from one
tower to the next, it is likely that there is heavy traffic along that highway.
• Of course, such a prediction is probabilistic and may not be accurate in each
circumstance, especially if there is a change in the underlying conditions.
However, with careful analysis and implementation, data that is not perfect can
be effectively employed to improve decision making in agriculture. For example, consider the large maize farmer who receives satellite maps of the fields for
which the farmer is responsible. Colors are used to identify conditions in the
field, with green indicating heavy vegetative growth. As one farmer reported to
the chapter’s author, an area marked in heavy green means either that the crop is
doing really well or that there is a heavy infestation of weeds. In either case, it is
worth the farmer’s time to physically investigate.
• In agriculture, as in most fields, descriptive efforts have been most common and
even those are relatively infrequent. However, within production agriculture,
knowing what has occurred – even if very accurately and precisely – may not
provide useful insights as to what should be done in the future.
• Production agriculture is complex, where biology, weather, and human actions
interact. Science-based methods have been employed to discern why crop and
livestock production occurs in the manner in which they do. Indeed, relative to
the big data topic, it might be useful to consider this as the “small data” process.
• The process starts with lab research employing the scientific method as a systematic process to gain knowledge through experimentation. Indeed, the scientific method is designed to ensure that the results of an experimental study did
not occur just by chance (Herren 2014). However, results left in the lab do not
lead to innovation and progress in the farm field. In the United States, the USDA,
Land Grant universities, and the private sector have collaborated to exploit
scientific advances. A highly effective, but distributed, system emerged where
knowledge gained in the laboratory was tested and refined on experimental plots
and then extended to agricultural producers.
8 Digital Technologies, Big Data, and Agricultural Innovation