Computer-implemented recommendation of side-by-side planting in agricultural fields
Techniques for recommending side-by-side plantings of pairs of hybrids or seeds include a server computer receiving agricultural data records that represent crop seed data describing seed and yield properties of hybrid seeds and first data for agricultural fields where the hybrid seeds were planted. The server receives second data for available hybrids and seeds and automatically calculates a dataset of success probability scores that describe the probability of a successful yield on the target fields. Data is organized as pairs to facilitate comparison of actual plantings to optimized plantings that have a probability of success (POS), in terms of yield lift or increased yield season-over-season, for different yield values. Confidence values are generated and stored in association with the POS values and can be used as a basis of visual output to support planting and/or field management decisions as part of an automated intelligent agricultural decision support system.
1. A computer-implemented method comprising:
receiving, over a digital data communication network at a server computer system, one or more agricultural data records of grower crop seed data describing seed and yield properties of one or more first products and first field geo-location data for one or more current agricultural fields where the one or more first products were planted, the one or more first products including hybrids or seeds;
receiving, over the digital data communication network at the server computer system, second crop seed data for available second products, which are different than the one or more first products;
generating, by the server computer system, and storing a plurality of different pairings of the one or more first products and the second products;
generating, by the server computer system, via a machine learning model, based on feature data for the plurality of different pairings, probability of success (POS) values each indicating a probability of a second product of one of the plurality of different pairings resulting in a crop yield greater than or equal to a crop yield of a first product of said one of the plurality of different pairings;
identifying each pairing of the plurality of different pairings for which the generated POS value is within a threshold of a POS value of a first one of the plurality of different pairings;
fitting a normal distribution to delta predicted yield values of the identified plurality of different pairings;
calculating a first area of the normal distribution that is greater than a first yield lift, the first yield lift being 0 bushels per acre, and the first area defining a first final POS value;
calculating a second area of the normal distribution that is greater than a second yield lift, the second yield lift being greater than 0 bushels per acre, and the second area defining a second final POS value; and
creating and causing displaying a visual graphical map on a computer display that identifies geographic locations of fields associated with the first area and/or the second area.
2. The computer-implemented method of claim 1 , further comprising selecting, from the grower crop seed data, the one or more first products based on a total difference of magnitude of relative maturity values of less than or equal to 3, prior to generating the plurality of different pairings; and
wherein generating the plurality of different pairings of the one or more first products and the second products includes generating the plurality of different pairings based on the selected one or more first products and the second products.
3. The computer-implemented method of claim 1 , wherein the machine learning model is programmed to classify each of the plurality of different pairings as correct when delta yield lift of said pairing, using best linear unbiased prediction (predicted yield) calculations, is greater than 5 bushels per acre.
4. The computer-implemented method of claim 1 , wherein the feature data comprises product POS difference values, previous year product predicted yield difference values, previous year standard error in difference of predicted yield estimation, and previous year product win rate for >0 bushels per acre.
5. The computer-implemented method of claim 1 , wherein the second yield lift includes one of: >2 bushels/acre and >5 bushels/acre.
6. The computer-implemented method of claim 1 , further comprising generating and causing displaying the first area and the second area, associated with the first yield lift and the second yield lift, respectively, in the visual graphical map.
7. The computer-implemented method of claim 1 , wherein creating and causing displaying the visual graphical map includes:
generating and causing displaying a plurality of visual graphical representations of boundaries of a plurality of different political units of a region of the map;
coloring each of the political units using a distinctive color that is associated with a particular different historic delta predicted yield value for fields in that political unit; and
generating and causing displaying distinctive graphical identifiers of the geographic locations of fields that were identified in the grower crop seed data for which the second products of the multiple different pairings resulted in a crop yield greater than or equal to a crop yield of the first product of the corresponding pairings.
8. A non-transitory computer-readable storage medium storing instructions which, when executed by one or more processors, cause the one or more processors to:
receive, over a digital data communication network at a server computer system, one or more agricultural data records of grower crop seed data describing seed and yield properties of one or more first products and first field geo-location data for one or more current agricultural fields where the one or more first products were planted, the one or more first products including hybrids or seeds;
receive, over the digital data communication network at the server computer system, second crop seed data for available second products, which are different than the one or more first products;
generate and store a plurality of different pairings of the one or more first products and the second products;
generate, via a machine learning model, based on feature data for the plurality of different pairings, probability of success (POS) values each indicative of a probability of a second product of one of the different pairings resulting in a crop yield greater than or equal to a crop yield of a first product of said one of the different pairings;
identify each pairing of the different pairings for which the generated POS value is within a threshold of a POS value of a first one of the different pairings;
fit a normal distribution to delta predicted yield values of the identifies different pairings;
calculate a first area of the normal distribution that is greater than a first yield lift, the first yield lift being 0 bushels per acre, and the first area defining a first final POS value;
calculate a second area of the normal distribution that is greater than a second yield lift, the second yield lift being greater than 0 bushels per acre, and the second area defining a second final POS value; and
create and cause to display a visual graphical map on a computer display that identifies geographic locations of fields associated with the first area and/or the second area.
9. The non-transitory computer-readable storage medium of claim 8 , wherein the instructions further cause the one or more processors to select, from the grower crop seed data, the one or more first products based on a total difference of magnitude of relative maturity values of less than or equal to 3, prior to generating the plurality of different pairings; and
wherein the instructions further cause the one or more processors, in generating the plurality of different pairings, to generate the plurality of different pairings based on the selected one or more first products and the second products.
10. The non-transitory computer-readable storage medium of claim 8 , wherein the machine learning model is programmed to classify each of the different pairings as correct when delta yield lift of said pairing, using best linear unbiased prediction (predicted yield) calculations, is greater than 5 bushels per acre.
11. The non-transitory computer-readable storage medium of claim 8 , wherein the feature data comprises product POS difference values, previous year product predicted yield difference values, previous year standard error in difference of predicted yield estimation, and previous year product win rate for >0 bushels per acre.
12. The non-transitory computer-readable storage medium of claim 8 , wherein the second yield lift includes one of: >2 bushels/acre and >5 bushels/acre.
13. The non-transitory computer-readable storage medium of claim 8 , wherein the instructions further cause the one or more processors to generate and cause to display the first area and the second area, associated with the first yield lift and the second yield lift, respectively, in the visual graphical map.
14. The non-transitory computer-readable storage medium claim 8 , wherein the instructions cause the one or more processors to, in creating and causing to display the visual graphical map on the computer display:
generate and cause to display a plurality of visual graphical representations of boundaries of a plurality of different political units of a region of the map;
color each of the political units using a distinctive color that is associated with a particular different historic delta predicted yield value for fields in that political unit; and
generate and cause to display distinctive graphical identifiers of the geographic locations of fields that were identified in the grower crop seed data for which the second products of the different pairings resulted in a crop yield greater than or equal to a crop yield of the first product of the corresponding pairings.