System for detection and analysis of viewership anomalies

Abstract

Aspects of the subject disclosure may include, for example, a method of receiving, by a processing system including a processor, viewership data for a plurality of viewers watching a program; identifying, by the processing system, an anomalous event from the viewership data, wherein the viewership data includes data points per viewer logged on a per second basis, and wherein the processing system identifies the anomalous event using a pattern recognition algorithm; analyzing, by the processing system, the viewership data to determine a reason for the anomalous event; and providing, by the processing system, a user interface that presents a comparison of the anomalous event versus a standard and indicia for the reason for the anomalous event. Other embodiments are disclosed.

What is claimed is: 
     
         1 . A device, comprising:
 a processing system including a processor; and   a memory that stores executable instructions that, when executed by the processing system, facilitate performance of operations, the operations comprising:
 receiving viewership data including tune in events and tune away events for viewers watching a presentation of media content, wherein the viewership data has a level of temporal granularity of a second or finer; 
 aggregating the viewership data for a plurality of viewers watching the presentation of the media content, thereby creating aggregated viewership data that identifies a number of viewers of the media content at the level of temporal granularity; 
 identifying an anomalous event from the aggregated viewership data, wherein the anomalous event comprises a change in the number of the viewers that differs from an expected change in the number of the viewers; and 
 adjusting the media content to mitigate the anomalous event. 
   
     
     
         2 . The device of  claim 1 , wherein the identifying further comprises:
 building an expected behavioral model of the plurality of viewers; and   comparing the aggregated viewership data with the expected behavioral model to determine the anomalous event.   
     
     
         3 . The device of  claim 2 , wherein the expected behavioral model is built by training a machine learning model on the aggregated viewership data. 
     
     
         4 . The device of  claim 1 , wherein the identifying determines the anomalous event using a time-series analysis algorithm. 
     
     
         5 . The device of  claim 1 , wherein the identifying determines the anomalous event using a pattern recognition algorithm. 
     
     
         6 . The device of  claim 1 , wherein the aggregated viewership data includes viewer demographics. 
     
     
         7 . The device of  claim 6 , wherein the viewership data includes tune-in events and tune-out events on a near real-time basis. 
     
     
         8 . The device of  claim 7 , wherein the operations further comprise determining a reason for the anomalous event. 
     
     
         9 . The device of  claim 1 , wherein the viewership data includes metadata derived from the media content. 
     
     
         10 . The device of  claim 9 , wherein the metadata identifies objects, faces, keywords, topics, or a combination thereof in the media content. 
     
     
         11 . The device of  claim 1 , wherein the processing system comprises a plurality of processors operating in a distributed computing environment. 
     
     
         12 . A non-transitory, machine-readable medium, comprising executable instructions that, when executed by a processing system including a processor, facilitate performance of operations, the operations comprising:
 receiving viewership data for viewers watching a presentation of a broadcast program, wherein the viewership data has a level of temporal granularity of a second or finer;   aggregating the viewership data for a plurality of viewers watching the broadcast program, thereby creating aggregated viewership data that identifies a number of the viewers of the broadcast program at the level of temporal granularity;   identifying an anomalous event from the aggregated viewership data, wherein the anomalous event comprises a change in the number of the viewers that differs from an expected change in the number of the viewers; and   adjusting the broadcast program to mitigate the anomalous event.   
     
     
         13 . The non-transitory, machine-readable medium of  claim 12 , wherein the anomalous event is associated with an advertisement displayed during the broadcast program. 
     
     
         14 . The non-transitory, machine-readable medium of  claim 12 , wherein the anomalous event is associated with a transition from the broadcast program to an advertisement displayed during the broadcast program. 
     
     
         15 . The non-transitory, machine-readable medium of  claim 14 , wherein the anomalous event comprises a tune away in the aggregated viewership data exceeding an expected tune away. 
     
     
         16 . The non-transitory, machine-readable medium of  claim 15 , wherein the adjusting uses a comparison of the tune away on a viewership demographic basis to mitigate the anomalous event. 
     
     
         17 . The non-transitory, machine-readable medium of  claim 12 , wherein the processing system comprises a plurality of processors operating in a distributed computing environment. 
     
     
         18 . A method, comprising:
 receiving, by a processing system including a processor, viewership data for a plurality of viewers watching a program, wherein the viewership data has a level of temporal granularity of a second or finer;   identifying, by the processing system, an anomalous event from the viewership data using a pattern recognition algorithm based on a change in a number of viewers that differs from an expected change in the number of the viewers of the program at the level of temporal granularity; and   adjusting, by the processing system, the program to mitigate the anomalous event.   
     
     
         19 . The method of  claim 18 , wherein the adjusting uses a machine learning model to provide feedback to producers of the program in real time. 
     
     
         20 . The method of  claim 18 , wherein the anomalous event is associated with an advertisement displayed during the program.