Methods and apparatus to identify co-relationships between media using social media
Methods, apparatus, systems and articles of manufacture are disclosed to identify co-relationships between media using social media. An example apparatus includes an audience estimator to: estimate a first audience of first media based on a first set of media-exposure social media messages corresponding to client devices referencing the first media, and estimate a second audience of second media based on a second set of media-exposure social media messages corresponding to the client devices referencing the second media. The example apparatus also includes a pairing classifier to: determine a pairing-score for a media-pairing based on the first and second audiences and the first and second sets of media-exposure social media messages, determine a relationship threshold to apply to the media-pairing based on the first media and the second media, and classify the media-pairing based on the pairing-score and the relationship threshold to improve an accuracy of a system associated with generating audience analysis information.
1. An apparatus comprising:
an audience estimator to:
estimate a first audience of first media based on a first set of media-exposure social media messages corresponding to client devices referencing the first media; and
estimate a second audience of second media based on a second set of media-exposure social media messages corresponding to the client devices referencing the second media; and
a pairing classifier to:
determine a pairing-score for a media-pairing based on the first and second audiences and the first and second sets of media-exposure social media messages;
determine a relationship threshold to apply to the media-pairing based on the first media and the second media; and
classify the media-pairing based on the pairing-score and the relationship threshold to improve an accuracy of a system associated with generating audience analysis information, the audience estimator and the paring classifier implemented by one or more logic circuits.
2. The apparatus as defined in claim 1 , further including a message logger to:
obtain the media-exposure social media messages corresponding to the client devices by querying a social media server based on media keywords and media posting timestamps; and
identify the media-pairing including the first media and the second media.
3. The apparatus as defined in claim 1 , wherein the audience estimator is to estimate the first audience by:
associating each of the media-exposure social media messages included in the first set with an author username; and
recording a first occurrence of each of the identified author usernames.
4. The apparatus as defined in claim 1 , wherein the audience estimator is to estimate the first audience by:
identifying an impression associated with each of the media-exposure social media messages included in the first set;
identifying a user identifier associated with the impression; and
recording a first occurrence of each of the identified impression user identifiers.
5. The apparatus as defined in claim 1 , wherein the audience estimator is to estimate the first audience by:
identifying a first subset of usernames associated with posting the first set of media-exposure social media messages to the social media server;
identifying a second subset of usernames associated with impressions of the media-exposure social media messages included in the first set, each of the usernames included in the second subset of usernames not present in the first subset of usernames.
6. The apparatus as defined in claim 1 , wherein the pairing classifier is to determine the pairing-score for the media-pairing by:
identifying a subset of audience members included in the first audience and the second audience; and
determining a size of the subset of audience members and a first count of media-exposure social media messages included in the first set.
7. The apparatus as defined in claim 1 , wherein the pairing classifier is to classify the media-pairing by:
applying the relationship threshold to the pairing-score; and
classifying the first media related to the second media when the pairing-score satisfies the relationship threshold.
8. A non-transitory computer readable medium comprising instructions that, when executed, cause a machine to at least:
estimate a first audience of first media based on a first set of media-exposure social media messages corresponding to client devices referencing the first media;
estimate a second audience of second media based on a second set of media-exposure social media messages corresponding to the client devices referencing the second media;
determine a pairing-score for a media-pairing based on the first and second audiences and the first and second sets of media-exposure social media messages;
determine a relationship threshold to apply to the media-pairing based on the first media and the second media; and
classify the media-pairing based on the pairing-score and the relationship threshold to improve an accuracy of a system associated with generating audience analysis information.
9. The non-transitory computer readable medium as defined in claim 8 , wherein the instructions are to further cause the machine to:
obtain the media-exposure social media messages corresponding to the client devices by querying a social media server based on media keywords and media posting timestamps; and
identify the media-pairing including the first media and the second media.
10. The non-transitory computer readable medium as defined in claim 8 , wherein the instructions are to further cause the machine to:
associate each of the media-exposure social media messages included in the first set with an author username; and
estimating the first audience by recording a first occurrence of each of the identified author usernames.
11. The non-transitory computer readable medium as defined in claim 8 , wherein the instructions are to further cause the machine to:
identify an impression associated with each of the media-exposure social media messages included in the first set;
identify a user identifier associated with the impression; and
estimate the first audience by recording a first occurrence of each of the identified impression user identifiers.
12. The non-transitory computer readable medium as defined in claim 8 , wherein the instructions are to further cause the machine to:
identify a first subset of usernames associated with posting the first set of media-exposure social media messages to the social media server; and
estimate the first audience by identifying a second subset of usernames associated with impressions of the media-exposure social media messages included in the first set, each of the usernames included in the second subset of usernames not present in the first subset of usernames.
13. The non-transitory computer readable medium as defined in claim 8 , wherein the instructions are to further cause the machine to:
identify a subset of audience members included in the first audience and the second audience; and
determine the pairing-score for the media-pairing by determining a size of the subset of audience members and a first count of media-exposure social media messages included in the first set.
14. The non-transitory computer readable medium as defined in claim 8 , wherein the instructions are to further cause the machine to:
apply the relationship threshold to the pairing-score; and
classify the first media related to the second media when the pairing-score satisfies the relationship threshold.
15. A method comprising:
estimating, by executing an instruction with a processor, a first audience of first media based on a first set of media-exposure social media messages corresponding to client devices referencing the first media;
estimating, by executing an instruction with the processor, a second audience of second media based on a second set of media-exposure social media messages corresponding to the client devices referencing the second media;
determining, by executing an instruction with the processor, a pairing-score for a media-pairing based on the first and second audiences and the first and second sets of media-exposure social media messages;
determining, by executing an instruction with the processor, a relationship threshold to apply to the media-pairing based on the first media and the second media; and
classifying, by executing an instruction with the processor, the media-pairing based on the pairing-score and the relationship threshold to improve an accuracy of a system associated with generating audience analysis information.
16. The method as defined in claim 15 , further including:
obtaining the media-exposure social media messages corresponding to the client devices by querying a social media server based on media keywords and media posting timestamps; and
identifying the media-pairing including the first media and the second media.
17. The method as defined in claim 15 , wherein the estimating of the first audience includes:
associating each of the media-exposure social media messages included in the first set with an author username; and
recording a first occurrence of each of the identified author usernames.
18. The method as defined in claim 15 , wherein the estimating of the first audience includes:
identifying an impression associated with each of the media-exposure social media messages included in the first set;
identifying a user identifier associated with the impression; and
recording a first occurrence of each of the identified impression user identifiers.
19. The method as defined in claim 15 , wherein the estimating of the first audience includes:
identifying a first subset of usernames associated with posting the first set of media-exposure social media messages to the social media server;
identifying a second subset of usernames associated with impressions of the media-exposure social media messages included in the first set, wherein each of the usernames included in the second subset of usernames is not present in the first subset of usernames.
20. The method as defined in claim 15 , wherein the determining of the pairing-score for the media-pairing includes:
identifying a subset of audience members included in the first audience and the second audience; and
determining a size of the subset of audience members and a first count of media-exposure social media messages included in the first set.