Popular media items data set with exponential decay

1. A method comprising:

computing, by a processing device, scores for a plurality of media items hosted by a content hosting platform based on a plurality of positive user actions associated with the plurality of media items;

ranking the plurality of media items based on the scores, the plurality of media items comprising one or more media items that are associated with the plurality of positive user actions and have not been featured by any platform external to the content hosting platform;

in view of the ranking and the one or more media items that are associated with the plurality of positive user actions and have not been featured by any platform external to the content hosting platform, selecting, from the plurality of media items hosted by the content hosting platform, highest ranked media items that have been featured by at least one platform external to the content hosting platform; and

creating a popular media item data set including the highest ranked media items that have been featured by the at least one platform external to the content hosting platform.

2. The method of claim 1 , wherein computing the scores for the plurality of media items further comprises:

determining the plurality of positive user actions associated with a media item;

combining a plurality of score contributions from the plurality of positive user actions associated with the media item to determine a value for a score for the media item; and

applying an exponential decay to the value for the score.

3. The method of claim 2 , wherein computing the scores for the plurality of media items further comprises:

adding the plurality of score contributions to a previous value for the score, wherein the determined plurality of positive user actions are positive user actions that occurred after the previous value was computed.

4. The method of claim 1 , wherein the plurality of positive user actions comprise a plurality of user views.

5. The method of claim 4 , wherein the plurality of user views comprise a first set of user views having durations that satisfy a first duration threshold and a second set of user views with durations that satisfy a second duration threshold, and wherein a first score contribution is applied to user views from the first set and a second score contribution is applied to user views from the second set.

6. The method of claim 4 , further comprising:

identifying, from the plurality of user views, subscriber views from subscribers to a media content source of the media item and non-subscriber views from non-subscribers to the media content source, wherein a first score contribution is applied to subscriber views and a second score contribution is applied to non-subscriber views.

7. The method of claim 2 , wherein the plurality of positive user actions further comprises at least one of a plurality of user indications that users like the media item, a plurality of user shares of the media item, or a plurality of user comments on the media item.

8. The method of claim 1 , further comprising:

determining authors of the plurality of media items; and

filtering a first media item of the plurality of media items from the popular media items data set if the first media item is authored by an author of a second higher ranked media item in the popular media items data set.

9. The method of claim 1 , wherein computing the scores for the plurality of media items comprises:

determining a user engagement value for a media item based on at least one of user shares of the media item, user indications of interest in the media item, user comments on the media item, user forwarding of the media item or playlists incorporating the media item;

determining a views value for the media item based on a number of user views of the media item;

determining a ratio between the views value and the user engagement value; and

filtering the media item from the popular media items data set responsive to determining that the ratio fails to satisfy a first threshold.

10. The method of claim 9 , further comprising:

boosting the score for the media item responsive to determining that the ratio satisfies a second threshold.

11. The method of claim 2 , wherein the applied exponential decay has a first exponential decay factor if an age of the media item is less than a threshold age and a second exponential decay factor if the age of the media item is greater than the threshold age.

12. The method of claim 1 , wherein computing the scores for the plurality of media items further comprises:

determining viewer demographics associated with a media item; and

responsive to determining that the viewer demographics associated with the media item vary from average demographics by more than a threshold value, reducing the score for the media item.

13. A non-transitory computer readable storage medium having instructions that, when executed by a processing device, cause the processing device to perform operations comprising:

computing, by the processing device, scores for a plurality of media items hosted by a content hosting platform based on a plurality of positive user actions associated with the plurality of media items;

ranking the plurality of media items based on the scores, the plurality of media items comprising one or more media items that are associated with the plurality of positive user actions and have not been featured by any platform external to the content hosting platform;

in view of the ranking and the one or more media items that are associated with the plurality of positive user actions and have not been featured by any platform external to the content hosting platform, selecting, from the plurality of media items hosted by the content hosting platform, highest ranked media items that have been featured by at least one platform external to the content hosting platform; and

creating a popular media item data set including the highest ranked media items that have been featured by the at least one platform external to the content hosting platform.

14. The computer readable storage medium of claim 13 , wherein computing the scores for the plurality of media items further comprises:

determining the plurality of positive user actions associated with a media item;

combining a plurality of score contributions from the plurality of positive user actions associated with the media item to determine a value for a score for the media item; and

applying an exponential decay to the value for the score.

15. The computer readable storage medium of claim 14 , wherein computing the scores for the plurality of media items further comprises:

adding the plurality of score contributions to a previous value for the score, wherein the determined plurality of positive user actions are those positive user actions that occurred after the previous value was computed.

16. The computer readable storage medium of claim 13 , the operations further comprising:

determining authors of the plurality of media items; and

filtering a first media item of the plurality of media items from the popular media items data set if the first media item is authored by an author of a second higher ranked media item in the popular media items data set.

17. The computer readable storage medium of claim 13 , wherein computing the scores for the plurality of media items further comprises:

determining a user engagement value for a media item based on at least one of user shares of the media item, user indications of interest in the media item, user comments on the media item, user forwarding of the media item or playlists incorporating the media item;

determining a views value for the media item based on a number of user views of the media item;

determining a ratio between the views value and the user engagement value; and

filtering the media item from the popular media items data set responsive to determining that the ratio fails to satisfy a first threshold.

18. A computing device comprising:

a memory; and

a processing device coupled to the memory, the processing device to:

compute scores for a plurality of media items hosted by a content hosting platform based on a plurality of positive user actions associated with the plurality of media items;

rank the plurality of media items based on the scores, the plurality of media items comprising one or more media items that are associated with the plurality of positive user actions and have not been featured by any platform external to the content hosting platform;

in view of the ranking and the one or more media items that are associated with the plurality of positive user actions and have not been featured by any source platform external to the content hosting platform, select, from the plurality of media items hosted by the content hosting platform, highest ranked media items that have been featured by at least one platform external to the content hosting platform; and

create a popular media item data set including the highest ranked media items that have been featured by the at least one platform external to the content hosting platform.

19. The computing device of claim 18 , wherein the plurality of media items comprise a plurality of videos, wherein the plurality of positive user actions comprise a first set of user views having durations that satisfy a first duration threshold and a second set of user views with durations that satisfy a second duration threshold, and wherein a first score contribution is applied to user views from the first set and a second score contribution is applied to user views from the second set.

20. The computing device of claim 18 , wherein computing the scores for the plurality of media items further comprises:

determining a user engagement value for a media item based on at least one of user shares of the media item, user indications of interest in the media item, user comments on the media item, user forwarding of the media item or playlists incorporating the media item;

determining a views value for the media item based on a number of user views of the media item;

determining a ratio between the views value and the user engagement value; and

filtering the media item from the popular media item data set responsive to determining that the ratio fails to satisfy a first threshold.