Systems and methods for predicting viewership and detecting anomalies

A system, comprising:

• at least one processor;
• at least one memory comprising instructions configured to be executed by the at least one processor to perform a method comprising:
• obtaining a first total amount of time that a first plurality of users is displayed a first set of content via a first plurality of content distributors during a plurality of past periods including a target period;
• obtaining a second total amount of time that a second plurality of users is displayed a second set of content via a second plurality of content distributors during the plurality of past periods including the target period;
• obtaining a first set of features associated with the display of the first set of content;
• obtaining a second set of features associated with the display of the second set of content;
• predicting, via a trained model, a predicted first total amount of time that the first plurality of users would be displayed the first set of content during the target period based on the obtained first set of features associated with the display of the first set of content and on the obtained first total amount of time that occurred during past periods that precede the target period;
• predicting, via the trained model, a predicted second total amount of time that the second plurality of users would be displayed the second set of content during the target period based on the obtained second set of features associated with the display of the second set of content and on the obtained second total amount of time that occurred during the past periods that precede the target period;
• comparing the obtained first total amount of time during the target period to the predicted first total amount of time and the obtained second total amount of time during the target period to the predicted second total amount of time;
• detecting an anomaly based on at least one of the obtained first total amount of time during the target period deviating from the predicted first total amount of time or the obtained second total amount of time during the target period deviating from the predicted second total amount of time;
• and
• alerting, based on the detection of the anomaly, to the anomaly.

A computer-implemented method for managing viewership data, the method comprising:

• obtaining a first total amount of time that a first plurality of users is displayed a first set of content via a first plurality of content distributors during a plurality of past periods including a target period;
• obtaining a second total amount of time that a second plurality of users is displayed a second set of content via a second plurality of content distributors during the plurality of past periods including the target period;
• obtaining a first set of features associated with the display of the first set of content;
• obtaining a second set of features associated with the display of the second set of content;
• predicting, via a trained model, a predicted first total amount of time that the first plurality of users would be displayed the first set of content during the target period based on the obtained first set of features associated with the display of the first set of content and on the obtained first total amount of time that occurred during past periods that precede the target period;
• predicting, via the trained model, a predicted second total amount of time that the second plurality of users would be displayed the second set of content during the target period based on the obtained second set of features associated with the display of the second set of content and on the obtained second total amount of time that occurred during the past periods that precede the target period;
• calculating a first ratio between the obtained first total amount and the predicted first total amount and a second ratio between the obtained second total amount and the predicted second total amount;
• determining that the first ratio or the second ratio is not within upper and lower thresholds;
• and
• alerting, based on the determination that the first ratio or the second ratio is not within the upper and lower thresholds, to an anomaly.