The Power of Predictive Analytics in Agriculture

Years ago if we would have been told computers, data, and technology would be scattered around every farm there may have been a push back. Today, as we look at adoption of technology, some paradigms continue to vary in what is widely accepted as a standard and what is “up and coming”.


One of the most exciting technologies presently being used and widely being transformed and developed has been the use of predictive analytics. Predictive analytics as a whole can be comprised of numerous different statistical abilities from modeling, machine learning, and data mining. Used for agriculture, these methods allow for analyzation of what has happened in the past on the farm, as well as what currently is happening and is going to happen, to make use of the data to predict the future and make decisions that impact the bottom line and end use of on-farm products.

Predictive analytics is not just a buzz word in agriculture currently, but a reality as actionable insights to make decisions on data and information to improve agronomic opportunities, such as timing of applications, product decisions, amounts of products, and profitability of decision making, and are all being utilized today with developments coming constantly. By learning from historical and future data based on measured variables, management and outcomes of decisions can more readily be made that can greatly impact efficiencies and processes. This is no easy task, as decisions and recommendations about the future require true datasets that have high confidence from field to field, even acre to acre, and within acre variability.

Predictive analytics require good data to be successful, and data that is incomplete or is incorrect will provide insights that are not fully analyzed. Data from in-field sensors, collection of input data at each level, and economic functions of decisions will continue to be critical for success of predictive analytics. As IoT rises and data collection becomes more essential to operations across the world, the power of making impactful, proactive, and profitable decisions that can increase opportunities and efficiencies across the farm continues to be of interest. This insight helps producers to make otherwise challenging agronomic decisions that can take time to reach the field every day quicker and easier. It gives them the opportunity to make a fast decision off of digital information, often with the ability to be unbiased to the source, but relied upon the facts. True agronomic knowledge is essential for success and the right outputs for each digital tool.


Finding opportunities for profit in down commodity prices is essential for profitability both in the short and long term. A small decision on a timing of an input application, for example, could mean the difference between profitability for that application. Predictive scores are given to each opportunity to help determine processes and decision making through analyzing datasets and confidence. These input parameters and variables for outcomes can all be addressed with predictive analytics.

Examples already being used in digital agriculture today range from market recommendations, pest modeling, soil test value, and crop yield predictions, as well as nutrient movement and behavior — all across varying conditions in and around each field. Wouldn’t it be nice to know the growth stage of the crop while having your morning coffee, before driving 20 miles to the field to make a fertigation decision? Or, knowing which fields soil test values have changed the most when budgets are tight and only half the farm can be sampled? Having high quality datasets and proper collection routes have made these possible already today, and the future continues to be bright. In-field validation of these decisions to ensure correct analysis is crucial to ensuring success. A sensor set incorrectly or a weather station not properly calibrated can lead to decision making that is skewed through analytics.

The future is bright for technology in agriculture, and the learning curves that have come in the last 10 years have been astounding. We are in an exciting time to continually enhance the industry and leave an impact on the world for years to come.

Leave a Reply

Avatar for igal aisenberg igal aisenberg says:

Predictive analytics based on on-going monitoring is part of the present and will certainly take a prominent role of future developments. With that unlike protected environments, farms are exposed to weather, animal, human and mechanical factors that require high maintenance, hamper the performance of in-field sensors and call for integrated systems that prioritize aerial monitoring. Without ignoring the value of on-the-ground sensing and satellites contribution, drones seem to have the best shot at filling the gap.

Avatar for Noel Magnin Noel Magnin says:

Nice article Scott. You just nailed it :”Predictive analytics require good data to be successful, and data that is incomplete or is incorrect will provide insights that are not fully analyzed”. This is the reason why Precision Agriculture is not so commonly embraced : if you don’t use the good data, you cannot take the good decisions and you do not get ROI. Most solution providers use the same data for completely different issues (yield, quality, pathogen detection …) and rely on analytics (AI, ML, NN …). No way this can work and too sad for farmers who are tech-believers. Technology can be high ROI if it is used wisely but most of the time it is not.

Avatar for Vicky Vicky says:

Such a nice informative post..