Generating and distributing digital surveys based on predicting survey responses to digital survey questions
The present disclosure relates to systems, methods, and non-transitory computer readable media for generating a predicted response to a digital survey question and identifying digital survey questions to remove from a digital survey. For example, the disclosed systems can reduce the number of digital survey questions distributed as part of a digital survey by identifying and removing similar digital survey questions. In addition, the disclosed systems can generate a predicted response to an unprovided digital survey question based on determining relationships between respondents. Further, based on respondent relationships, the disclosed systems can identify digital survey questions that a respondent is likely to answer and can provide the digital survey questions to a respondent device of the respondent.
1. A system comprising:
at least one processor; and
a non-transitory computer-readable medium comprising instructions that, when executed by the at least one processor, cause the system to:
receive, from a respondent device, responses to a subset of answered digital survey questions from among a plurality of digital survey questions;
determine, based on the responses to the subset of answered digital survey questions, a relationship between a respondent associated with the respondent device and at least one other respondent of a plurality of respondents; and
generate, for the respondent, a predicted response to an unanswered digital survey question from among the plurality of digital survey questions based on the relationship between the respondent and the at least one other respondent.
2. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to generate a survey matrix comprising indications of, for the plurality of respondents, respective responses to digital survey questions from the plurality of digital survey questions.
3. The system of claim 2 , wherein the instructions cause the system to determine the relationship between the respondent and the at least one other respondent by:
generating, based on the survey matrix, a respondent vector associated with the respondent and at least one other respondent vector associated with the at least one other respondent; and
determining a distance between the respondent vector and the at least one other respondent vector in a vector space.
4. The system of claim 3 , wherein the instructions cause the system to generate the predicted response by:
identifying, within the vector space, a second respondent vector that is closest to the respondent vector associated with the respondent; and
selecting, as the predicted response, a response given by a second respondent associated with the second respondent vector.
5. The system of claim 3 , wherein the instructions cause the system to determine the relationship between the respondent and the at least one other respondent by further:
identifying respondent attributes associated with the respondent and the at least one other respondent; and
comparing the respondent attributes associated with the respondent with respondent attributes associated with the at least one other respondent.
6. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to generate the predicted response to the unanswered digital survey question by generating a predicted response to a digital survey question not provided to the respondent device.
7. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to select, from the plurality of digital survey questions, a digital survey question to provide to the respondent device by:
determining a relevance score for the digital survey question in relation to the respondent;
determining that the relevance score satisfies a relevance threshold; and
based on the relevance score satisfying the relevance threshold, selecting the digital survey question to provide to the respondent device.
8. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to:
generate a plurality of digital survey question vectors corresponding to the plurality of digital survey questions;
identify, from the plurality of digital survey question vectors, a first digital survey question vector and a second digital survey question vector that are within a threshold similarity of each other; and
based on identifying the first digital survey question vector and the second digital survey question within the threshold similarity of each other:
select a first digital survey question associated with the first digital survey question vector to provide to the respondent device; and
exclude, from providing to the respondent device, a second digital survey question associated with the second digital survey question vector.
9. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to:
generate, based on the relationship between the respondent and the at least one other respondent, a predicted digital survey question that satisfies a threshold relevance with respect to the respondent; and
provide the predicted digital survey question to the respondent device.
10. A non-transitory computer-readable medium comprising instructions that, when executed by at least one processor, cause a computer device to:
receive, from a respondent device, responses to a subset of answered digital survey questions from among a plurality of digital survey questions;
determine, based on the responses to the subset of answered digital survey questions, a relationship between a respondent associated with the respondent device and at least one other respondent of a plurality of respondents; and
generate, for the respondent, a predicted response to an unanswered digital survey question from among the plurality of digital survey questions based on the relationship between the respondent and the at least one other respondent.
11. The non-transitory computer readable medium of claim 10 , further comprising instructions that, when executed by the at least one processor, cause the computer device to generate a survey matrix comprising indications of, for the plurality of respondents, respective responses to digital survey questions from the plurality of digital survey questions.
12. The non-transitory computer readable medium of claim 11 , wherein the instructions cause the computer device to determine the relationship between the respondent and the at least one other respondent by:
generating, based on the survey matrix, a respondent vector associated with the respondent and at least one other respondent vector associated with the at least one other respondent; and
determining a distance between the respondent vector and the at least one other respondent vector in a vector space.
13. The non-transitory computer readable medium of claim 12 , wherein the instructions cause the computer device to generate the predicted response by:
identifying, within the vector space, a second respondent vector that is closest to the respondent vector associated with the respondent; and
selecting, as the predicted response, a response given by a second respondent associated with the second respondent vector.
14. The non-transitory computer readable medium of claim 12 , wherein the instructions cause the computer device to determine the relationship between the respondent and the at least one other respondent by further:
identifying respondent attributes associated with the respondent and the at least one other respondent; and
comparing the respondent attributes associated with the respondent with respondent attributes associated with the at least one other respondent.
15. The non-transitory computer readable medium of claim 10 , further comprising instructions that, when executed by the at least one processor, cause the computer device to select, from the plurality of digital survey questions, a digital survey question to provide to the respondent device by:
determining a relevance score for the digital survey question in relation to the respondent;
determining that the relevance score satisfies a relevance threshold; and
based on the relevance score satisfying the relevance threshold, selecting the digital survey question to provide to the respondent device.
16. A method comprising:
receiving, from a respondent device, responses to a subset of answered digital survey questions from among a plurality of digital survey questions;
determining, based on the responses to the subset of answered digital survey questions, a relationship between a respondent associated with the respondent device and at least one other respondent of a plurality of respondents; and
generating, for the respondent, a predicted response to an unanswered digital survey question from among the plurality of digital survey questions based on the relationship between the respondent and the at least one other respondent.
17. The method of claim 16 , further comprising generating a survey matrix comprising indications of, for the plurality of respondents, respective responses to digital survey questions from the plurality of digital survey questions.
18. The method of claim 17 , wherein determining the relationship between the respondent and the at least one other respondent comprises:
generating, based on the survey matrix, a respondent vector associated with the respondent and at least one other respondent vector associated with the at least one other respondent; and
determining a distance between the respondent vector and the at least one other respondent vector in a vector space.
19. The method of claim 18 , wherein generating the predicted response comprises:
identifying, within the vector space, a second respondent vector that is closest to the respondent vector associated with the respondent; and
selecting, as the predicted response, a response given by a second respondent associated with the second respondent vector.
20. The method of claim 16 , further comprising:
generating a plurality of digital survey question vectors corresponding to the plurality of digital survey questions;
identifying, from the plurality of digital survey question vectors, a first digital survey question vector and a second digital survey question vector that are within a threshold similarity of each other; and
based on identifying the first digital survey question vector and the second digital survey question within the threshold similarity of each other:
selecting a first digital survey question associated with the first digital survey question vector to provide to the respondent device; and
excluding, from providing to the respondent device, a second digital survey question associated with the second digital survey question vector.