The Evolution of digital tools in forecasting crop productivity and their importance in recommending Crop Inputs

This March 6th grain completed one year in Brazil, but with work that has been begun in 2020 with UPL and Sinagro, a distributor of crop inputs with more than 30 stores in the Brazilian cerrado.

After this year of introduction of the company in Brazil, the platform has 1.5 million hectares mapped, with 10 thousand areas registered, in 2500 farmers, and more than 170 thousand images captured and processed that are helping us more and more with data and information.

In this digital game, information is the gold mine, the more information, the more assertive the solutions.

As a consultant and curious about digital solutions, I wanted to share with you some of the lessons learned from this startup, which has a platform to classify crops on the ground through remote sensing data, time series analysis, and ML/DL (machine learning/deep learning) algorithms.

The insight generated by the platform is being extremely valuable for estimating the real market size, for greater assertiveness in production estimates and expansion strategies for distributors, as well as to help agronomists and specialists in the recommendation of crop inputs.

In strong partnership with some of the largest Brazilian agribusiness companies such as Syngenta Seeds, Sinagro, UPL, Adubos Araguaia, DNAgro, IZagro, and CESB, some work fronts were developed, such as:

1 - Crop Mapping, where the number of hectares planted with Soybean and Second Season Corn in the Brazilian Cerrado was mapped.

2 - The yield curves where yields were estimated and compared these yields with previous years and with areas of neighboring farms and even other regions.

3 - And a Pilot to assist in the recommendation of Biostimulants and demand generation to observe how effective or not was the application of the product to promoting yield increase.

One of the reasons that I appreciate promoting 6th Grain in Brazil was the technological capacity of its founders Vladimir Eskin and Molly Brown, in this report, you will have some success stories, but also improvements, to have more and more digitization as an important part in the decision-making of farmers, consultants, and companies.

1: The Crop Mapping was an initiative of Syngenta Seeds where the objective was to map the area planted with soybeans by municipalities and the area afterward with the second crop of corn.

In Brazil, fortunately, we managed to have in almost all our planted areas the possibility of making a second crop, and Syngenta wanted to know how much of the soybean area was already using this technique.

It took 6 months of work to identify areas without a single man in the field whereas 6th Grain developed a system based on remote sensing that allowed the quantitative analysis of cultivated areas with a resolution of 10m. This helped us to understand, in addition to the expansion of the second crop, the market potential of each region, to allocate more or fewer resources and estimate growth zones for the coming years.

The integration of scientific algorithms in development platforms for business decision-making generated information with an assertiveness superior to 98% compared to other sources of information such as Conab and IBGE. Without the need for so much manpower, the costs for these estimates were much more adequate to the reality of agriculture companies.

In the figures below, we can see where the soybean areas were in green and where the second season corn areas were in orange.

When the two areas were intersected, we had the fields in Mato Grosso (purple) that were used for soybean production during the first harvest followed by corn production during the second harvest.

Another application of this tool, which was not used in Brazil but in a country in Africa, also by Syngenta, was the analysis of the overlapping of the cultivation maps with the allocation of the infrastructure where through images, we could allocate the places for generating demands, the crop inputs stores, salespeople, storage locations, etc., resulting in opportunities for expansion and optimization of the resources.

Today I see the frantic movement to expand new stores to places that are already extremely saturated, and this tool could help a lot in better decision making.

Being with 6th grain is allowing me to help my clients in various decision making such as:

• Estimate the market potential in $/Ha for Crop inputs, Seeds, silo structures by region/sales territory

• Estimate market share and penetration estimates in specific regions

• Identify the most suitable regions for market expansion or acquisition of retail locations

• Optimize sales force efforts

• Determine the most suitable locations for demand generation and development testing

• Determining the best moments for the sales campaign according to the stage of Crop growth, temperatures, and homogeneity

• Identification of Active Customers and  Non-Customers

2 – The Yield Curves

This work was developed with the Crop Dealer called Adubos Araguaia where 30 farmers were assisted, and their productivity was monitored in real-time.

The digital solution showed a very high assertiveness, and it was just not better because the historical data was not good enough, the results showed that the more information we have from previous years, the greater the assertiveness of the curve.

In the image below you can see the areas and how much were the prediction and the real results

This information allowed Adubos Araguaia to monitor the productivity of its customers in real-time and thus make decisions regarding trivial recommendations such as whether to apply a booster dose of fungicide, application of foliar nutrition, or a biostimulant and to monitor possible risks of non-production. and consequently, any break in the delivery of production or payment of bills.

The Tool also allowed Araguaia technicians to compare the areas of their producers with areas of their neighbors, showing the superiority or not of their programs.

In the image below we have a comparison of the area of a client of Adubos Araguaia with productivity estimates of 3.6 T/ha compared to an area of a non-client with 3.3 T/ha in neighboring regions.

What I more liked about this pilot was the possibility of comparison and continuous learning between agronomic techniques since the action and reaction of the actions taken can be verified almost in real-time.

3 - Demand Generation and Biostimulant Recommendation

And the third pilot that I followed was the demand generation allied to the correct moments of application of Biostimulants.

Companies invest a lot of money in people, products, and infrastructure to make field days, demonstrative areas, and experimental fields, and the results are often discredited by farmers.

With technology, this discredit can be even greater because companies, through NDVI images, can choose the most fertile areas or areas with fewer problems to place their best genetic material or product.

This work was carried out for DNagro, a foliar nutrition, and biostimulants company located in the city of Ourinhos in the state of São Paulo. All demonstrative and comparative areas were plotted on the platform and the evolution was monitored as shown in the figure below.

In the Figure below we can see the orange areas being the areas where the DNA products were applied and in the blue areas the standard treatment of the farmer.

With this data in hand, the performance of one area can be compared daily with another in different locations, all within easy access for the sales, development, and marketing team.

In the figure below we have the two estimated yield curves placed side by side and you can see the moments when the application of the product really made a difference or not.

All these monitored curves can greatly help agronomists to apply the product at the right time. In the example below, we see the application of the product in a time of water stress and the resumption of culture soon after the application of the product.

What I found most interesting was the possibility that we as agronomists have a very good precision to know if we are going in the right direction or not towards great yields.

The summary that I make of these tools is that in a short time we will have digital solutions that will increasingly allow us to achieve higher productivity and manage more and more the costs and the needs or not of the applications of crop inputs and that we also have a lot of points to improve.

But it's like I say to everyone in my consultancies "Who doesn't measure, doesn't control" and with the database that these tools are storing, our recommendations will be increasingly assertive.

One thing I can guarantee we will have a bright future in agriculture with this digitization. Greater productivity, Greater Assertiveness, and Greater sustainability due to the correct and accurate use of crop inputs through greater assertiveness in our recommendations.