Systems and methods for regional demand estimation

1 . A computer-implemented system for demand estimation of a region, the system comprising:

at least one memory storing instructions; and

at least one processor configured to execute the instructions to perform operations comprising:

receive, via a first user device, user input configuring a future trigger event or time interval for generating a demand estimation request;

construct a configuration file including the future trigger event or the time interval;

automatically generate a demand estimation request in response to determining an occurrence of the future trigger event or the time interval indicated in the configuration file;

retrieve an item identifier associated with an item for demand estimation of a region in response to the generation of the demand estimation request;

access overall demand forecast data for the item associated with the retrieved item identifier using a first machine learning model trained on past order data associated with the item, wherein accessing the overall demand forecast data comprises communicating with a warehouse management system to obtain live item order fulfillment updates through mobile devices;

identify one or more geographical regions at a uniform granularity level, wherein granularity level is determined based on item orders served by centers in the one or more geographical regions;

evaluate a demand share estimate of the item associated with the retrieved item identifier in each of the identified one or more geographical regions, wherein the demand share estimate of the item is based on retrieving order data associated with the item and the retrieved order data is limited due to the item being at least one of: a seasonal item, a new item, or a slow-selling item;

determine a set of item identifiers associated with a segment of items related to the item associated with the retrieved item identifier, wherein the segment of items are determined based on being a similar type as the item;

evaluate a demand estimation for the segment of items associated with the set of item identifiers for each of the identified one or more geographical regions using a second machine learning model trained on past order data associated with the segment of items;

generate demand estimation of the item associated with the retrieved item identifier at one or more geographical regions at the uniform granularity level using a Bayesian framework, wherein the Bayesian framework is provided as input demand share of the item and demand estimation of the segment of items in a geographical region, and overall demand forecast data for the item;

generate demand estimation of the item associated with the retrieved item identifier at one or more higher levels of granularity, wherein the demand estimation of the item at the one or more higher levels of granularity is generated by aggregating demand estimate of the item at the one or more geographical regions;

provide the generated demand estimation to a supply chain management system to generate one or more purchase orders to purchase and stock a sufficient quantity to satisfy forecasted demand for the item associated with the retrieved item identifier;

generate, on a second user device, an instruction to stow or retrieve the item associated with the retrieved item identifier identifying a location;

detect a scan of a barcode associated with the item associated with the retrieved item identifier using the second user device;

prompt a scan of a barcode at the identified location; and

send data to the warehouse management system indicating that the item associated with the retrieved item identifier has been stowed at the identified location.

2 . The computer-implemented system of claim 1 , wherein accessing the overall demand forecast further comprises:

retrieving past order data associated with the item from an order database, wherein a selected item is associated with an order if a version of the selected item is ordered; and

wherein the first machine learning model is trained on the retrieved past order data.

3 . The computer-implemented system of claim 1 , wherein identifying the one or more geographical regions at uniform granularity level is based on similarity of at least of one of:

size of the one or more geographical regions;

number of orders served in the one or more geographical regions; or

geographic areas served by one common delivery center.

4 . The computer-implemented system of claim 1 , wherein the demand share estimate of the item associated with the retrieved item identifier is adjusted for seasonal demands.

5 . The computer-implemented system of claim 1 , wherein the demand estimation for the segment of items includes point estimates.

6 . The computer-implemented system of claim 1 , wherein determining a set of item identifiers associated with a segment of items related to the item associated with the retrieved item identifier is performed on a just-in-time basis.

7 . The computer-implemented system of claim 1 ,

wherein the segment of items are related to each other, wherein a relationship between the segment of items is based on order combinations including the segment of items; and

wherein the segment of items overlaps with a second segment of items partially or fully as a subset of the second segment of items.

8 . The computer-implemented system of claim 1 , wherein the response is structured in a tabular manner comprising the generated demand estimate and demand estimates for smaller regions within the region.

9 . The computer-implemented system of claim 1 , wherein aggregating demand estimate into demand estimates of the item associated with the retrieved item identifier at the one or more geographic regions is defined in the configuration file.

10 . The computer-implemented system of claim 1 , wherein aggregating demand estimate of the item associated with the retrieved item identifier at the one or more geographical regions is based on a type of the item.

11 . The computer-implemented system of claim 1 , wherein aggregating demand estimate of the item associated with the retrieved item identifier at the one or more geographical regions further comprises:

determining uncertainty of the demand estimation of the item; and

aggregating the uncertainty of the demand estimation of the item at the one or more geographical regions.

12 . The computer-implemented system of claim 11 , wherein determining uncertainty of the demand estimation of the item associated with the retrieved item identifier further comprises:

computing the uncertainty of the demand estimation of the segment of items associated with the set of item identifiers;

computing uncertainty of overall demand forecast for the item; and

determining the uncertainty of the demand estimation of the item based on the uncertainty of the demand estimation of the of items, the uncertainty of the overall demand forecast for the item, and the demand share estimate of the item.

13 . The computer-implemented system of claim 1 ,

wherein retrieving an item identifier associated with a selected item for demand estimation further comprises transforming data in the configuration file into the item identifier.

14 . The computer-implemented system of claim 1 ,

wherein the at least one processor is further configured to:

receive, via the user device, user input indicating at least one rule to determine a segment of items related to the item associated with the retrieved item identifier,

wherein the configuration file is constructed to include the at least one rule.

15 . The computer-implemented system of claim 1 ,

wherein the at least one processor is further configured to:

transform the generated demand estimation into a tabular form indicating demand estimation of the region and all smaller regions encompassed by the region; and

wherein the response to the retrieved item identifier comprises the transformed demand estimation.

16 . The computer-implemented system of claim 1 , further comprising:

tracking, via a front end system, user search data;

transmitting a query to the front end system requesting the user search data;

parsing the user search data to determine the segment of items related to the item associated with the retrieved item identifier.

17 . A computer-implemented method for inventory demand estimation of a region, the method comprising:

receiving, via a first user device, user input configuring a future trigger event or time interval for generating a demand estimation request;

constructing a configuration file including the future trigger event or the time interval;

automatically generating a demand estimation request in response to determining an occurrence of the future trigger event or the time interval indicated in the configuration file;

retrieving an item identifier associated with an item for demand estimation of a region in response to the generation of the demand estimation request;

accessing overall demand forecast data for the item associated with the retrieved item identifier using a first machine learning model trained on past order data associated with the item, wherein accessing the overall demand forecast data comprises communicating with a warehouse management system to obtain live item order fulfillment updates through mobile devices;

identifying one or more geographical regions at a uniform granularity level, wherein granularity level is determined based on item orders served by centers in the one or more geographical regions;

evaluating a demand share estimate of the item associated with the retrieved item identifier in each of the identified one or more geographical regions, wherein the demand share estimate of the item is based on retrieving order data associated with the item and the retrieved order data is limited due to the item being at least one of: a seasonal item, a new item, or a slow-selling item;

determining a set of item identifiers associated with a segment of items related to the item associated with the retrieved item identifier, wherein the segment of items are determined based on being a similar type as the item;

evaluating a demand estimation for the segment of items associated with the set of item identifiers for each of the identified one or more geographical regions using a second machine learning model trained on past order data associated with the segment of items;

generating demand estimation of the item associated with the retrieved item identifier at one or more geographical regions at the uniform granularity level using a Bayesian framework, wherein the Bayesian framework is provided as input demand share of the item and demand estimation of the segment of items in a geographical region, and overall demand forecast data for the item;

generating demand estimation of the item associated with the retrieved item identifier at one or more higher levels of granularity, wherein the demand estimation of the item at the one or more higher levels of granularity is generated by aggregating demand estimate of the item at the one or more geographical regions;

provide the generated demand estimation to a supply chain management system to generate one or more purchase orders to purchase and stock a sufficient quantity to satisfy forecasted demand for the item associated with the retrieved item identifier;

generate, on a second user device, an instruction to stow or retrieve the item associated with the retrieved item identifier identifying a location;

detect a scan of a barcode associated with the item associated with the retrieved item identifier using the second user device;

prompt a scan of a barcode at the identified location; and

send data to the warehouse management system indicating that the item associated with the retrieved item identifier has been stowed at the identified location.

18 . The method of claim 17 , wherein accessing the overall demand forecast further comprises:

retrieving past order data associated with the item associated with the retrieved item identifier from an order database, wherein a selected item is associated with an order if a version of the selected item is ordered; and

wherein the first machine learning model is trained on the retrieved past order data.

19 . The method of claim 17 , wherein identifying the one or more geographical regions at uniform granularity level is based on similarity of at least of one of:

size of the one or more geographical regions;

number of orders served in the one or more geographical regions; or

geographic areas served by one common delivery center.

20 . The method of claim 17 ,

wherein the segment of items are related to each other, wherein a relationship between the segment of items is based on order combinations including the segment of items; and

wherein the segment of items overlaps with a second segment of items partially or fully as a subset of the second segment of items.

21 . The method of claim 17 , wherein the response is structured in a tabular manner comprising the generated demand estimate and demand estimates for smaller regions within the region.

22 . The method of claim 17 , wherein aggregating demand estimate into demand estimates of the item associated with the retrieved item identifier at the one or more geographic regions is based on type of the item.

23 . The method of claim 17 , wherein aggregating demand estimate into demand estimates of the item associated with the retrieved item identifier at the one or more geographic regions further comprises:

determining uncertainty of the demand estimation of the item, wherein the uncertainty of the demand estimation of the item is based on the uncertainty of the demand estimation of the segment of items, uncertainty of overall demand forecast for the item, and the demand share estimate of the item; and

aggregating the uncertainty of the demand estimation of the item at the one or more geographical regions.

24 . A computer-implemented system for low-latency aggregated-data provision, the system comprising:

at least one memory storing instructions; and

at least one processor configured to execute the instructions to perform operations comprising:

receive, via a first user device, user input configuring a future trigger event or time interval for generating a demand estimation request;

construct a configuration file including the future trigger event or the time interval;

automatically generate a demand estimation request in response to determining an occurrence of the future trigger event or the time interval indicated in the configuration file;

retrieve an item identifier associated with an item for demand estimation of a region in response to the generation of the demand estimation request;

access overall demand forecast data for the item associated with the retrieved item identifier using a first machine learning model trained on past order data associated with the item, wherein accessing the overall demand forecast data comprises communicating with a warehouse management system to obtain live item order fulfillment updates through mobile devices;

identify one or more geographical regions at a uniform granularity level, wherein granularity level is determined based on item orders served by centers in the one or more geographical regions;

evaluate a demand share estimate of the item associated with the retrieved item identifier in each of the identified one or more geographical regions, wherein the demand share estimate of the item is based on retrieving order data associated with the item and the retrieved order data is limited due to the item being at least one of: a seasonal item, a new item, or a slow-selling item;

determine a set of item identifiers associated with a segment of items related to the item associated with the retrieved item identifier, wherein the segment of items are determined based on being a similar type as the item;

evaluate a demand estimation for the segment of items associated with the set of item identifiers for each of the identified one or more geographical regions using a second machine learning model trained on past order data associated with the segment of items;

generate demand estimation of the item associated with the retrieved item identifier at one or more geographical regions at the uniform granularity level using a Bayesian framework, wherein the Bayesian framework is provided as input demand share of the item and demand estimation of the segment of items in a geographical region, and overall demand forecast data for the item;

generate demand estimation of the item associated with the retrieved item identifier at one or more higher levels of granularity, wherein the demand estimation of the item at the one or more higher levels of granularity is generated by aggregating demand estimate of the item at the one or more geographical regions;

provide the generated demand estimation to a supply chain management system to generate one or more purchase orders to purchase and stock a sufficient quantity to satisfy forecasted demand for the item associated with the retrieved item identifier;

generate, on a second user device, an instruction to stow or retrieve the item associated with the retrieved item identifier identifying a location;

detect a scan of a barcode associated with the item associated with the retrieved item identifier using the second user device;

prompt a scan of a barcode at the identified location; and

send data to the warehouse management system indicating that the item associated with the retrieved item identifier has been stowed at the identified location.