Consumer insights analysis by identifying a similarity in public sentiments for a pair of entities
In one embodiment, a method includes receiving a request to identify a similarity in public sentiments for each pair from a plurality of entities from a second computing device, where the request includes names of the plurality of entities, accessing a table of word vector relationships, where the table of word vector relationships includes a plurality of unique n-grams and their corresponding word vectors, and where each of the word vectors represents a semantic context of a corresponding n-gram as a point in a d-dimensional embedding space, looking up word vectors corresponding to each of the names using the table, calculating, for each of the word vectors, a similarity metric to each of the word vectors, and sending a response message to the second computing device, where the response message includes calculated similarity metrics corresponding to all the pairs of the word vectors.
1. A method comprising:
by a first computing device in an online social network, receiving, from a second computing device, a request to identify a similarity in public sentiments of a particular demographic for each pair from a plurality of commercial brands, wherein the request comprises brand names of the plurality of commercial brands, and wherein the request comprises one or more conditions characterizing the particular demographic;
by the first computing device, collecting content objects that are generated by users of the online social network who satisfy the one or more conditions;
by the first computing device, constructing a corpus of text as training data for a word embedding model by extracting text from the collected content objects generated by the users of the online social network who satisfy the one or more conditions;
by the first computing device, constructing a table of word vector relationships by training the word embedding model using the constructed corpus of text as training data, wherein the table of word vector relationships comprises a plurality of unique n-grams and their corresponding word vectors, wherein each of the word vectors represents a semantic context of a corresponding n-gram as a point in a d-dimensional embedding space;
by the first computing device, looking up, using the table, word vectors corresponding to each of the brand names;
by the first computing device, calculating, for each of the word vectors, a similarity metric to each of the word vectors; and
by the first computing device, sending, to the second computing device, a response message, wherein the response message comprises calculated similarity metrics corresponding to all the pairs of the word vectors.
2. The method of claim 1 , wherein the plurality of unique n-grams in the table are selected from the corpus of text.
3. The method of claim 1 , wherein the word embedding model is a word2vec model.
4. The method of claim 1 , wherein the looking up the word vectors comprises looking up each of the plurality of brand names.
5. The method of claim 1 , wherein the similarity metric is a cosine similarity, a Euclidean distance, or a Jaccard similarity coefficient.
6. The method of claim 1 , wherein the response message comprises instructions to display the calculated similarity metrics.
7. The method of claim 6 , wherein the calculated similarity metrics are color-coded, wherein a color represents any number within a pre-determined range.
8. The method of claim 1 , wherein the content objects were created within a pre-determined period of time.
9. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
receive, from a second computing device, a request to identify a similarity in public sentiments of a particular demographic for each pair from a plurality of commercial brands, wherein the request comprises brand names of the plurality of commercial brands, and wherein the request comprises one or more conditions characterizing the particular demographic;
collect content objects that are generated by users of the online social network who satisfy the one or more conditions;
construct a corpus of text as training data for a word embedding model by extracting text from the collected content objects generated by the users of the online social network who satisfy the one or more conditions;
construct a table of word vector relationships by training the word embedding model using the constructed corpus of text as training data, wherein the table of word vector relationships comprises a plurality of unique n-grams and their corresponding word vectors, wherein each of the word vectors represents a semantic context of a corresponding n-gram as a point in a d-dimensional embedding space;
look up, using the table, word vectors corresponding to each of the brand names;
calculate, for each of the word vectors, a similarity metric to each of the word vectors;
and send, to the second computing device, a response message, wherein the response message comprises calculated similarity metrics corresponding to all the pairs of the word vectors.
10. The media of claim 9 , wherein the plurality of unique n-grams in the table are selected from the corpus of text.
11. The media of claim 9 , wherein the word embedding model is a word2vec model.
12. The media of claim 9 , wherein the looking up the word vectors comprises looking up each of the plurality of brand names.
13. The media of claim 9 , wherein the similarity metric is a cosine similarity, a Euclidean distance, or a Jaccard similarity coefficient.
14. The media of claim 9 , wherein the response message comprises instructions to display the calculated similarity metrics.
15. The media of claim 14 , wherein the calculated similarity metrics are color-coded, wherein a color represents any number within a pre-determined range.
16. A system comprising:
one or more processors; and
one or more computer-readable non-transitory storage media coupled to one or more of the processors and comprising instructions operable when executed by one or more of the processors to cause the system to:
receive, from a second computing device, a request to identify a similarity in public sentiments of a particular demographic for each pair from a plurality of commercial brands, wherein the request comprises brand names of the plurality of commercial brands, and wherein the request comprises one or more conditions characterizing the particular demographic;
collect content objects that are generated by users of the online social network who satisfy the one or more conditions;
construct a corpus of text as training data for a word embedding model by extracting text from the collected content objects generated by the users of the online social network who satisfy the one or more conditions;
construct a table of word vector relationships by training the word embedding model using the constructed corpus of text as training data, wherein the table of word vector relationships comprises a plurality of unique n-grams and their corresponding word vectors, wherein each of the word vectors represents a semantic context of a corresponding n-gram as a point in a d-dimensional embedding space;
look up, using the table, word vectors corresponding to each of the brand names;
calculate, for each of the word vectors, a similarity metric to each of the word vectors; and
send, to the second computing device, a response message, wherein the response message comprises calculated similarity metrics corresponding to all the pairs of the word vectors.