Methods and apparatus to identify non-traditional asset-bundles for purchasing groups using social media
Methods, apparatus, systems and articles of manufacture are disclosed to identify non-traditional asset-bundles for purchasing groups using social media. An example method includes identifying an asset-bundle in a social media message and generating a profile for a user associated with the social media message. The example method also includes identifying a plurality of social media messages posted by cohorts of the user based on the generated profile and classifying the asset-bundle based on occurrences of the asset-bundle in the plurality of social media messages.
1. A method comprising:
sending, by executing an instruction with a processor, a first network communication from an audience measurement entity server to a social media server, the first network communication requesting social media messages;
parsing, by executing an instruction with the processor at the audience measurement entity server, a second network communication received from the social media server to identify a social media message that references a first asset;
analyzing, by executing an instruction with the processor, the social media message using image recognition to identify a second asset in the social media message;
associating, by executing an instruction with the processor, the first asset and the second asset in an asset-bundle;
determining, by executing an instruction with the processor, if the asset-bundle has been previously identified to reduce the amount of processing requirements of a central facility;
storing, by executing an instruction with the processor, the asset-bundle in memory of a central facility in response to determining that the asset-bundle had not been previously identified;
generating, by executing an instruction with the processor, a profile for a user associated with the social media message;
determining, by executing an instruction with the processor, if the profile for the user is already stored in memory of a central facility;
in response to determining that the profile for the user is already stored in memory of the central facility, foregoing storage of the generated profile of the user to reduce the amount of storage requirements of the central facility;
identifying, by executing an instruction with the processor, a plurality of social media messages posted by cohorts of the user based on the generated profile; and
classifying, by executing an instruction with the processor, the asset-bundle based on occurrences of the asset-bundle in the plurality of social media messages, the classification to be stored in a memory of the audience measurement entity server.
2. A method as defined in claim 1 , wherein generating the profile for the user comprises:
analyzing a second plurality of social media messages for characteristic elements, each of the second plurality of social media messages associated with the user; and
applying statistical methods to the characteristic elements to generate the profile for the user.
3. A method as defined in claim 2 , wherein the statistical methods includes applying Bayesian analysis.
4. A method as defined in claim 1 , wherein identifying the plurality of social media messages posted by cohorts of the user comprises:
identifying profiles that share a characteristic element included in the profile; and
requesting social media messages posted by users associated with the identified profiles.
5. A method as defined in claim 1 , wherein classifying the asset-bundle comprises:
determining a count for the asset-bundle based on a number of occurrences of the asset bundle in the plurality of social media messages;
comparing the count for the asset-bundle to a first threshold; and
classifying the asset-bundle a coincidental pairing when the count for the asset-bundle does not satisfy the first threshold.
6. A method as defined in claim 5 further comprising:
comparing the count to a second threshold when the count satisfies the first threshold; and
classifying the asset-bundle a traditional pairing when the count for the asset-bundle satisfies the second threshold.
7. A method as defined in claim 6 further comprising classifying the asset-bundle as a non-traditional pairing when the count for the asset-bundle does not satisfy the second threshold.
8. A method as defined in claim 5 further comprising:
when the count for the asset-bundle satisfies the first threshold, identifying a first category for a first asset of the asset-bundle;
identifying a second category for a second asset of the asset-bundle; and
classifying the asset-bundle as a traditional pairing when the first category is related to the second category.
9. A method as defined in claim 8 further comprising classifying the asset-bundle as a non-traditional pairing when the first category is not related to the second category.
10. An apparatus comprising:
a query generator to send a first network communication from an audience measurement entity server to a social media server, the first network communication requesting social media messages;
an asset identifier to:
parse a second network communication received from the social media server, to identify a social media message that references a first asset;
analyze the social media message using image recognition to identify a second asset in the social media message;
associate the first asset and the second asset in an asset-bundle; and
determine if the asset-bundle has been previously identified to reduce the amount of processing requirements of a central facility;
store the asset-bundle in memory of a central facility in response to determining that the asset-bundle had not been previously identified;
a profile generator to:
generate a profile for a user associated with the social media message; and
determine if the profile for the user is already stored in memory of a central facility;
in response to determining that the profile for the user is already stored in memory of the central facility, foregoing storage of the generated profile of the user to reduce the amount of storage requirements of the central facility; and
a bundle classifier to (1) identify a plurality of social media messages posted by cohorts of the user based on the generated profile and (2) classify the asset-bundle based on occurrences of the asset-bundle in the plurality of social media messages, the classification to be stored in a memory of the audience measurement entity server, the query generator, the asset identifier, the profile generator, and the bundle classifier to be implemented by a logic circuit.
11. An apparatus as defined in claim 10 , wherein the profile generator is to:
analyze a second plurality of social media messages for characteristic elements, each of the second plurality of social media messages associated with the user; and
apply Bayesian analysis to the characteristic elements to generate the profile for the user.
12. An apparatus as defined in claim 10 , wherein the bundle classifier is to identify the plurality of social media messages posted by cohorts of the user by:
identifying profiles that share a characteristic element included in the profile; and
requesting social media messages posted by users associated with the identified profiles.
13. An apparatus as defined in claim 10 , wherein the bundle classifier is to classify the asset-bundle by:
determining a count for the asset-bundle based on a number of occurrences of the asset bundle in the plurality of social media messages;
comparing the count for the asset-bundle to a first threshold; and
classifying the asset-bundle a coincidental pairing when the count for the asset-bundle does not satisfy the first threshold.
14. An apparatus as defined in claim 13 , wherein the bundle classifier is to:
compare the count to a second threshold when the count satisfies the first threshold; and
classify the asset-bundle a traditional pairing based on a determination that the count for the asset-bundle satisfies the second threshold.
15. An apparatus as defined in claim 14 , wherein the bundle classifier is to classify the asset-bundle as a non-traditional pairing based on a determination that the count for the asset-bundle does not satisfy the second threshold.
16. An apparatus as defined in claim 13 , wherein the bundle classifier is to:
identify a first category for a first asset of the asset-bundle when the count for the asset-bundle satisfies the first threshold;
identify a second category for a second asset of the asset-bundle; and
classify the asset-bundle as a traditional pairing based on a determination that the first category is related to the second category.
17. An apparatus as defined in claim 16 , wherein the bundle classifier is to classify the asset-bundle as a non-traditional pairing based on a determination that the first category is not related to the second category.
18. A non-transitory tangible machine readable storage medium comprising instructions that, when executed, cause a processor to at least:
send a first network communication from an audience measurement entity server to a social media server, the first network communication requesting social media messages;
parse a second network communication received from the social media server, to identify a social media message that references a first asset;
analyze the social media message using image recognition to identify a second asset in the social media message;
associate the first asset and the second asset in an asset-bundle;
determine if the asset-bundle has been previously identified to reduce the amount of processing requirements of a central facility;
store the asset-bundle in memory of a central facility in response to determining that the asset-bundle had not been previously identified;
generate a profile for a user associated with the social media message;
determine if the profile for the user is already stored in memory of a central facility;
in response to determining that the profile for the user is already stored in memory of the central facility, foregoing storage of the generated profile of the user to reduce the amount of storage requirements of the central facility; and
identify a plurality of social media messages posted by cohorts of the user based on the generated profile; and
classify the asset-bundle based on occurrences of the asset-bundle in the plurality of social media messages, the classification to be stored in a memory of the audience measurement entity server.
19. A non-transitory tangible machine readable storage medium as defined in claim 18 , wherein the instructions are further to cause the processor to generate the profile for the user by:
analyzing a second plurality of social media messages for characteristic elements, each of the second plurality of social media messages associated with the user; and
applying statistical methods to the characteristic elements to generate the profile for the user.
20. A non-transitory tangible machine readable storage medium as defined in claim 19 , wherein the instructions are further to cause the processor to apply Bayesian analysis to the characteristic elements.
21. A non-transitory tangible machine readable storage medium as defined in claim 18 , wherein the instructions are further to cause the processor to identify the plurality of social media messages posted by cohorts of the user by:
identifying profiles that share a characteristic element included in the profile; and
requesting social media messages posted by users associated with the identified profiles.
22. A non-transitory tangible machine readable storage medium as defined in claim 18 , wherein the instructions are further to cause the processor to classify the asset-bundle by:
determining a count for the asset-bundle based on a number of occurrences of the asset bundle in the plurality of social media messages;
comparing the count for the asset-bundle to a first threshold; and
classifying the asset-bundle a coincidental pairing based on a determination that the count for the asset-bundle does not satisfy the first threshold.
23. A non-transitory tangible machine readable storage medium as defined in claim 22 , wherein the instructions are further to cause the processor to:
compare the count to a second threshold when the count satisfies the first threshold; and
classify the asset-bundle a traditional pairing based on a determination that the count for the asset-bundle satisfies the second threshold.
24. A non-transitory tangible machine readable storage medium as defined in claim 23 , wherein the instructions are further to cause the processor to classify the asset-bundle as a non-traditional pairing based on a determination that the count for the asset-bundle does not satisfy the second threshold.
25. A non-transitory tangible machine readable storage medium as defined in claim 22 , wherein the instructions are further to cause the processor to:
identify a first category for a first asset of the asset-bundle when the count for the asset-bundle satisfies the first threshold;
identify a second category for a second asset of the asset-bundle; and
classify the asset-bundle as a traditional pairing based on a determination that the first category is related to the second category.
26. A non-transitory tangible machine readable storage medium as defined in claim 25 , wherein the instructions are further to cause the processor to classify the asset-bundle as a non-traditional pairing based on a determination that the first category is not related to the second category.