Network traffic forecasting for non-ticketed events

• US 10,477,416 B2
• Filed: 10/13/2017
• Issued: 11/12/2019
• Est. Priority Date: 10/13/2017
• Status: Expired due to Fees

• Current Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Bhorkar, Abhijeet, Jana, Shraboni
• Primary Examiner(s): Mattis, Jason E
• Application Number:
  US15/783,789
• Publication Number:
  US 20190116505A1
• Time in Patent Office: 760 Days
• CPC Class Codes: G06N 3/02 Neural networks
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: H04L 47/70 Admission control; Resource allocation
• CPC Class Codes: H04W 24/02 Arrangements for optimising operational condition
• CPC Class Codes: H04W 24/08 Testing, supervising or monitoring using real traffic
• CPC Class Codes: H04W 24/10 Scheduling measurement reports ; Arrangements for measurement reports
• CPC Class Codes: H04W 84/042 Public Land Mobile systems, e.g. cellular systems

Abstract

An example method includes estimating a number of people expected to attend a non-ticketed event, based on electronic data including social media postings, estimating a per-person amount of network traffic expected to be generated on a communications network during the non-ticketed event, based at least in part on historical per-person network traffic statistics for a historical event of a similar type to the non-ticketed event, calculating an amount of total network traffic expected to be generated during the non-ticketed event, based at least on the number of people expected to attend the non-ticketed event and the per-person amount of network traffic, and implementing a modification to an infrastructure of the communications network in a geographic location of the non-ticketed event, based at least in part on the amount of total network traffic expected to be generated during the non-ticketed event.

Claims

1. A method comprising:

estimating a number of people expected to attend a non-ticketed event, based on electronic data including social media postings, wherein the estimating the number of people expected to attend the non-ticketed event comprises;

generating a popularity score for the non-ticketed event, based at least in part on the social media postings, wherein the popularity score is calculated as a number of mentions of the non-ticketed event in the social media postings within a defined window of time divided by a number of mentions of the non-ticketed event in the social media postings that originate at a geographic location of the non-ticketed event; and

calculating the number of people expected to attend the non-ticketed event as the popularity score multiplied by a maximum capacity of the non-ticketed event;

estimating a per-person amount of network traffic expected to be generated on a communications network during the non-ticketed event, based at least in part on historical per-person network traffic statistics for a historical event of a similar type to the non-ticketed event;

calculating an amount of total network traffic expected to be generated during the non-ticketed event, based at least on the number of people expected to attend the non-ticketed event and the per-person amount of network traffic; and

implementing a modification to an infrastructure of the communications network in the geographic location of the non-ticketed event, based at least in part on the amount of total network traffic expected to be generated during the non-ticketed event.

2. The method of claim 1, wherein the non-ticketed event is a planned event that is set to occur at a future time.

3. The method of claim 1, wherein the non-ticketed event is a planned event that is currently ongoing.

4. The method of claim 1, wherein the non-ticketed event is an unplanned event that is currently ongoing.

5. The method of claim 1, wherein the number of people expected to attend the non-ticketed event is updated based on real-time estimates of road traffic within a defined geographic area surrounding in a vicinity of the non-ticketed event.

6. The method of claim 1, wherein the estimating the per-person amount of network traffic comprises performing a time series estimate using the historical per-person network traffic statistics for the historical event.

7. The method of claim 1, wherein the per-person amount of network traffic is calculated as a function of time as (T(t)+S(t)+R(t)), where S(t) is a seasonality component, T(t) is a trend component as a function of time, and R(t) represents a measure of error, wherein the trend component assumes an exponential growth in the per-person amount of network traffic, and wherein the seasonality component is modeled to consider traffic variation over weekdays.

8. The method of claim 1, wherein the estimating the per-person amount of network traffic comprises using a deep recurrent neural network to model seasonality, burst, and trend components of the historical per-person network traffic statistics for the historical event.

9. The method of claim 1, wherein the modification includes deploying temporary infrastructure and equipment to accommodate the amount of total network traffic expected to be generated during the non-ticketed event.

10. The method of claim 9, wherein the temporary infrastructure and equipment includes a mobile evolved nodeB.

11. The method of claim 9, wherein the temporary infrastructure and equipment includes a distributed antenna system.

12. The method of claim 9, wherein the temporary infrastructure and equipment includes a portable base station.

13. The method of claim 1, wherein the modification includes instantiating a virtual network function.

14. The method of claim 1, wherein the historical event is identified by performing a linear discriminant analysis on the electronic data in order to classify the non-ticketed event.

15. A system comprising:

a processor; and

a computer-readable medium storing instructions which, when executed by the processor, cause the processor to perform operations, the operations comprising;

estimating a number of people expected to attend a non-ticketed event, based on electronic data including social media postings, wherein the estimating the number of people expected to attend the non-ticketed event comprises;

generating a popularity score for the non-ticketed event, based at least in part on the social media postings, wherein the popularity score is calculated as a number of mentions of the non-ticketed event in the social, media postings within a defined window of time divided by a number of mentions of the non-ticketed event in the social media postings that originate at a geographic location of the non-ticketed event; and

calculating the number of people expected to attend the non-ticketed event as the popularity score multiplied by a maximum capacity of the non-ticketed event;

estimating a per-person amount of network traffic expected to be generated on a communications network during the non-ticketed event, based at least in part on historical per-person network traffic statistics for a historical event of a similar type to the non-ticketed event;

calculating an amount of total network traffic expected to be generated during the non-ticketed event, based at least on the number of people expected to attend the non-ticketed event and the per-person amount of network traffic; and

implementing a modification to an infrastructure of the communications network in the geographic location of the non-ticketed event, based at least in part on the amount of total network traffic expected to be generated during the non-ticketed event.

16. The system of claim 15, wherein the non-ticketed event is a planned event that is set to occur at a future time.

17. The system of claim 15, wherein the non-ticketed event is a planned event that is currently ongoing.

18. The system of claim 15, wherein the non-ticketed event is an unplanned event that is currently ongoing.

19. The system of claim 15, wherein the number of people expected to attend the non-ticketed event is updated based on real-time estimates of road traffic within a defined geographic area in a vicinity of the non-ticketed event.

20. A non-transitory computer-readable medium storing instructions which, when executed by a processor, cause the processor to perform operations, the operations comprising:

estimating a number of people expected to attend a non-ticketed event, based on electronic data including social media postings, wherein the estimating the number of people expected to attend the non-ticketed event comprises;

generating a popularity score for the non-ticketed event, based at least in part on the social media postings, wherein the popularity score is calculated as a number of mentions of the non-ticketed event in the social media postings within a defined window of time divided by a number of mentions of the non-ticketed event in the social media postings that originate at a geographic location of the non-ticketed event; and

calculating the number of people expected to attend the non-ticketed event as the popularity score multiplied by a maximum capacity of the non-ticketed event;

estimating a per-person amount of network traffic expected to be generated on a communications network during the non-ticketed event, based at least in part on historical per-person network traffic statistics for a historical event of a similar type to the non-ticketed event;

calculating an amount of total network traffic expected to be generated during the non-ticketed event, based at least on the number of people expected to attend the non-ticketed event and the per-person amount of network traffic; and

implementing a modification to an infrastructure of the communications network in the geographic location of the non-ticketed event, based at least in part on the amount of total network traffic expected to be generated during the non-ticketed event.