Satellite-based agricultural modeling
An online agricultural system manages and optimizes interactions of entities within the system to enable the execution of transaction and the transportation of crop products. The online agricultural system accesses historic and environmental data describing factors that may impact crop product transactions and/or transportation to determine market prices for crop products and crop product transportation. Responsive to receiving a request from an entity, the online agricultural system determines an optimal transaction for the entity, such as a price for selling a crop product, an available crop product for purchase, or a transportation opportunity to transport a crop product.
1. A method comprising:
generating a first training set of data comprising historic crop information and current crop information representative of a first crop product;
training a first machine-learned model configured to predict a future characteristic of the first crop product using the training set of data;
generating a second training set of data comprising remote sensor data corresponding to the crop product type of the first crop product and associated historic quality specification data corresponding to the crop product type of the first crop product;
training a second machine-learned model configured to predict a quality specification for the first crop product based on remote sensor data corresponding to the first crop product using the second training set of data;
for each of a plurality of crop producers:
receiving a first request to list a first crop product within an online agricultural system, the first request identifying a first location of the first crop product and a crop product type of the first crop product;
applying the first machine-learned model to the first location of the first crop product and the crop product type to predict a first future characteristic of the first crop product;
modifying a first user interface presented to the crop producer to include the predicted first future characteristic, the user interface requesting one or both of a first crop product price and a first quantity of the first crop product; and
determining a first quality specification for the first crop product listed within the first request by applying the second machine-learned model to remote sensor data corresponding to the first crop product; and
for each of a plurality of prospective acquiring entities:
receiving a second request to acquire a second crop product, the second request identifying a second quality requirement of the second crop product and at least the crop product type of the second crop product, a second quantity of the second crop product, a second crop product price, and a second location to which the second crop product is to be delivered;
monitoring, by a server in real-time, the first quality specification of each first request received from a crop producer to identify a set of first requests with first quality specifications that satisfy the second quality requirement of the second crop product identified by the second request; and
modifying a second user interface presented to the prospective acquiring entity to include for one or more first requests in the identified set of first requests: the corresponding predicted first future characteristic, the corresponding first crop product price, the corresponding first quantity of the first crop product, the corresponding first location of the first crop product, and the corresponding first quality specification of the first crop product.
2. The method of claim 1 , wherein the first location of the first crop product comprises one of: a field boundary, a production location of the first crop product, and a storage location of the first crop product.
3. The method of claim 1 , further comprising, for each of the plurality of crop producers, obtaining environmental data for the first location of the first crop product, and wherein the environmental data comprises current or historic weather data and/or one or more soil characteristics.
4. The method of claim 1 , wherein, for at least one of the plurality of crop producers, the first crop product is a crop that has not been harvested, and wherein one or more of the type of the first crop product and the first quantity of the first crop product is inferred from the remote sensing data.
5. The method of claim 1 , wherein the second interface is modified to further include an expected distribution of prices or an expected average price of the second crop product for the prospective acquiring entity to acquire the second quantity of the second crop product.
6. The method of claim 1 , wherein the second interface is modified to further include a distribution of geographic locations from which the second crop product is expected to be acquired for the prospective acquiring entity to acquire the second quantity of the second crop product.
7. The method of claim 1 , wherein the remote sensor data corresponding to the first crop product is obtained from remote sensors in real-time during transportation of the first crop product.
8. The method of claim 1 , wherein the first location of the first crop product is the production location of the first crop product and the remote sensor data is satellite data representative of the production location of the first crop product.
9. The method of claim 1 , wherein the remote sensor data is obtained from a set of remote sensors including one or more of: GPS sensors, in-cargo sensors, hyperspectral sensors, NIR or visible spectroscopy sensors, temperature sensors, moisture sensors, humidity sensors, sensors to detect a presence of pests, and CO2 level sensors.
10. A non-transitory computer-readable storage medium storing executable instructions that, when executed, cause an online agricultural system to perform steps comprising:
generating a first training set of data comprising historic crop information and current crop information representative of a first crop product;
training a first machine-learned model configured to predict a future characteristic of the first crop product using the training set of data;
generating a second training set of data comprising remote sensor data corresponding to the crop product type of the first crop product and associated historic quality specification data corresponding to the crop product type of the first crop product;
training a second machine-learned model configured to predict a quality specification for the first crop product based on remote sensor data corresponding to the first crop product using the second training set of data;
for each of a plurality of crop producers:
receiving a first request to list a first crop product within the online agricultural system, the first request identifying a first location of the first crop product and a crop product type of the first crop product;
applying the first machine-learned model to the first location of the first crop product and the crop product type to predict a first future characteristic of the first crop product;
modifying a first user interface presented to the crop producer to include the predicted first future characteristic, the user interface requesting one or both of a first crop product price and a first quantity of the first crop product; and
determining a first quality specification for the first crop product listed within the first request by applying the second machine-learned model to remote sensor data corresponding to the first crop product; and
for each of a plurality of prospective acquiring entities:
receiving a second request to acquire a second crop product, the second request identifying a second quality requirement of the second crop product and at least the crop product type of the second crop product, a second quantity of the second crop product, a second crop product price, and a second location to which the second crop product is to be delivered;
monitoring, by a server in real-time, the first quality specification of each first request received from a crop producer to identify a set of first requests with first quality specifications that satisfy the second quality requirement of the second crop product identified by the second request; and
modifying a second user interface presented to the prospective acquiring entity to include for one or more first requests in the identified set of first requests: the corresponding predicted first future characteristic, the corresponding first crop product price, the corresponding first quantity of the first crop product, the corresponding first location of the first crop product, and the corresponding first quality specification of the first crop product.
11. The computer-readable storage medium of claim 10 , wherein the first location of the first crop product comprises one of: a field boundary, a production location of the first crop product, and a storage location of the first crop product.
12. The computer-readable storage medium of claim 10 , further comprising, for each of the plurality of crop producers, obtaining environmental data for the first location of the first crop product, and wherein the environmental data comprises current or historic weather data and/or one or more soil characteristics.
13. The computer-readable storage medium of claim 10 , wherein, for at least one of the plurality of crop producers, the first crop product is a crop that has not been harvested, and wherein one or more of the type of the first crop product and the first quantity of the first crop product is inferred from the remote sensing data.
14. The computer-readable storage medium of claim 10 , wherein the second interface is modified to further include an expected distribution of prices or an expected average price of the second crop product for the prospective acquiring entity to acquire the second quantity of the second crop product.
15. The computer-readable storage medium of claim 10 , wherein the second interface is modified to further include a distribution of geographic locations from which the second crop product is expected to be acquired for the prospective acquiring entity to acquire the second quantity of the second crop product.
16. The computer-readable storage medium of claim 10 , wherein the remote sensor data corresponding to the first crop product is obtained from remote sensors in real-time during transportation of the first crop product.
17. The computer-readable storage medium of claim 10 , wherein the first location of the first crop product is the production location of the first crop product and the remote sensor data is satellite data representative of the production location of the first crop product.
18. The computer-readable storage medium of claim 10 , wherein the remote sensor data is obtained from a set of remote sensors including one or more of: GPS sensors, in-cargo sensors, hyperspectral sensors, NIR or visible spectroscopy sensors, temperature sensors, moisture sensors, humidity sensors, sensors to detect a presence of pests, and CO2 level sensors.
19. An online agricultural system comprising:
a non-transitory computer-readable storage medium storing executable instructions that, when executed, cause the online agricultural system to perform steps comprising:
generating a first training set of data comprising historic crop information and current crop information representative of a first crop product;
training a first machine-learned model configured to predict a future characteristic of the first crop product using the training set of data;
generating a second training set of data comprising remote sensor data corresponding to the crop product type of the first crop product and associated historic quality specification data corresponding to the crop product type of the first crop product;
training a second machine-learned model configured to predict a quality specification for the first crop product based on remote sensor data corresponding to the first crop product using the second training set of data;
for each of a plurality of crop producers:
receiving a first request to list a first crop product within the online agricultural system, the first request identifying a first location of the first crop product and a crop product type of the first crop product;
applying the first machine-learned model to the first location of the first crop product and the crop product type to predict a first future characteristic of the first crop product;
modifying a first user interface presented to the crop producer to include the predicted first future characteristic, the user interface requesting one or both of a first crop product price and a first quantity of the first crop product; and
determining a first quality specification for the first crop product listed within the first request by applying the second machine-learned model to remote sensor data corresponding to the first crop product; and
for each of a plurality of prospective acquiring entities;
receiving a second request to acquire a second crop product, the second request identifying a second quality requirement of the second crop product and at least the crop product type of the second crop product, a second quantity of the second crop product, a second crop product price, and a second location to which the second crop product is to be delivered;
monitoring, by a server in real-time, the first quality specification of each first request received from a crop producer to identify a set of first requests with first quality specifications that satisfy the second quality requirement of the second crop product identified by the second request; and
modifying a second user interface presented to the prospective acquiring entity to include for one or more first requests in the identified set of first requests: the corresponding predicted first future characteristic, the corresponding first crop product price, the corresponding first quantity of the first crop product, the corresponding first location of the first crop product, and the corresponding first quality specification of the first crop product.
20. A method comprising:
generating a first training set of data comprising historic crop information and current crop information representative of a first crop product;
training a first machine-learned model configured to predict a future characteristic of the first crop product using the training set of data;
for each of a plurality of crop producers:
receiving a first request to list a first crop product within an online agricultural system, the first request identifying a first location of the first crop product and a crop product type of the first crop product;
applying the first machine-learned model to the first location of the first crop product and the crop product type to predict a first future characteristic of the first crop product; and
modifying a first user interface presented to the crop producer to include the predicted first future characteristic, the user interface requesting one or both of a first crop product price and a first quantity of the first crop product; and
for each of a plurality of prospective acquiring entities;
receiving a second request to acquire a second crop product, the second request identifying a second quality requirement of the second crop product and at least the crop product type of the second crop product, a second quantity of the second crop product, a second crop product price, and a second location to which the second crop product is to be delivered;
monitoring, by a server in real-time, a first quality specification of each first request received from a crop producer to identify a set of first requests with first quality specifications that satisfy the second quality requirement of the second crop product identified by the second request; and
modifying a second user interface presented to the prospective acquiring entity to include for one or more first requests in the identified set of first requests: the corresponding predicted first future characteristic, the corresponding first crop product price, the corresponding first quantity of the first crop product, the corresponding first location of the first crop product, and the corresponding first quality specification of the first crop product.
21. The method of claim 1 , wherein the first quality specification is determined in real-time in response to detecting a change in remote sensor data corresponding to the first crop product.
22. The method of claim 1 , wherein the first machine-learned model and the second machine-learned model each comprise one or more of: a neural network, a k-means clustering model, a reinforcement learning model, a linear support vector machine, an AdaBoost model, a logistics regression model, a naïve Bayes classifier, a random forest model, a bagged tree model, a decision tree model, a boosted tree model, and a boosted stump model.
23. The method of claim 1 , wherein the first machine-learned model comprises a first neural network, and wherein the second machine-learned model comprises a second neural network different from the first neural network.