By Tamar Rosati - Head of Granular / President of Digital Business at Corteva Agriscience.
With the rise of smart sensors, connected equipment and other agricultural applications of Internet of things (IoT) technology, it’s estimated that the average farm generates approximately 500,000 data points per day. Multiply that by the number of U.S. farms and you’ve already arrived at a staggering 1 trillion new pieces of farm data generated each and every day.
But as an industry, ag tech is still evolving how we harvest that data and turn it into insights farmers can use during the growing season to stay ahead of the curve. I think of this evolution in data analysis as a progression from simple, to smart, to predictive.
We’ve heard directly from farmers just that. In fact, one midwest farmer shared that he knows the future of digitally-supported farming lies in predictive analytics.
In his words, “Putting all of the data I collect on the farm to work has been a struggle over the years, but now I am getting mid-season predictions of my yield that enable me to make strategic decisions that truly impact my bottom line - not just for this year, but I anticipate for upcoming years as well”.
Most ag software to date has focused mainly on capturing data and turning it into simple analytics, which often take the form of digitized spreadsheets and provide only a retrospective look at outcomes that have already come to pass. What was my yield on this field? Which hybrids performed the best? How did I do against my budget? There is undoubtedly value in this level of analytics, especially when viewed over time.
Moving data from spreadsheets to software, however, begins to unlock the power of computer models to advance into smart analytics, which can pull out trends and patterns that would be impossible to observe with human analysis and spreadsheets alone. This is where multiple data sets begin to be paired with one another to tease out relationships, such as the relationship between moisture level and yield.
For example, the Directed Scouting feature in Granular Insights uses satellite imagery paired with other data to track plant growth throughout the season and flag fields that are showing the greatest variability in crop health or greatest crop health change in season, directing customers to the highest risk fields to take action before it’s too late.
Predictive analytics take that same data one step further by forecasting possible outcomes before they happen. Going back to our example, once the fields with potential crop health issues have been identified, predictive analytics can then forecast yields and how they might change based on how those issues are addressed, even recommending optimal harvest timing based on all the data flowing into the computer model.
This is where ag tech companies can deliver the most value to farmers, unlocking insights that are only possible with software and predictive models. As an example, having actionable, mid-season predictions of yield allows for strategic decision making on things like grain storage, marketing plans and maximizing inputs, all of which help farmers head into harvest with confidence.
In Granular Insights, this happens within Yield Predict, a feature that gives farmers a prediction of their yields a couple months prior to harvest. As we prepare to take Yield Predict from a beta product to a broader launch for the 2023 season with several key enhancements, we’re cognizant of the fact that many farmers have been overwhelmed and under-impressed with promises of what data can do for them.
That’s why we’ve been transparent about Yield Predict, giving farmers clear information as to how the tool generates predictions and reminding them that predictive analytics is not an exact science. There will always be a margin of error on forecasts generated by computer models. Human judgement and agronomic expertise from farmers and their trusted advisors will always be required to decide the best course of action based on the information at hand. What predictive analytics can provide are more informed courses of action, which ultimately lead to more confident decision making. That’s when farmers unlock the true value of data.