WebQ is being retired as of June 16, 2022! After that date, I do not expect this package will be functional.
An R interface to the University of Washington WebQ REST API for Online Surveys
rwebq pulls data from the UW Catalyst WebQ survey API and formats it in easy-to-use R tibbles.
This project was created for my own use is not maintained by UW IT. It therefore does not support all of the functionality of the WebQ REST API, only the features I needed for my own work. Feel free to reach out and/or submit a PR if you would like to collaborate on extending this package.
NOTE: Catalyst WebQ is available for members of University of Washington only.
You can install this package from Github using devtools::install_github('lpiep/rwebq').
Next, you must generate a private key using your UW login. Once you have done so, store your credentials using the webq_config function. This function stores your netid and API key as environment variables (CATALYST_NETID, CATALYST_KEY).
webq_config(netid = 'sue', key = 'abcdefghijklmnopqrstuvwxyz123456789')
You'll need to find the survey_id for your survey in order to pull data from it. You can list all of the surveys you have access to with webq_surveys().
> webq_surveys()
# A tibble: 3 x 3
name survey_id response_count
<chr> <chr> <dbl>
1 My Cool Survey 123456 176
2 My Cooler Survey 654321 117
3 My Less Cool Survey 010101 86
You can also list the participant_ids, along with the times when they took the survey. The function also returns the rest_id, which is included in the API response, but does not appear to be populated (at least for my surveys).
> webq_participants('123456')
# A tibble: 3 x 5
survey_id participant_id rest_id start_date end_date
<chr> <chr> <list> <chr> <chr>
1 123456 20595146 <chr [0]> 2021-05-04 08:45:48 2021-05-04 09:02:15
2 123456 19349250 <chr [0]> 2019-12-17 10:25:05 2019-12-17 10:31:18
3 123456 18334462 <chr [0]> 2018-11-14 17:46:59 2018-11-14 23:56:27
And, of course, you can pull the responses to your survey. The API supports filtering responses by participant_id, and so you can provide a character vector of participant_ids to the webq_responses function. If you don't provide this vector, the function returns responses for all participants. The API does not support filtering by question_id, but you can filter the results afterwards.
> webq_responses('123456') # We didn't provide `participant_ids`, so all participant's responses will be returend.
# A tibble: 3 x 6
question_id question_type question_content text_values numeric_values participant_id
<chr> <chr> <chr> <chr> <chr> <chr>
1 3657118 Multiple choice - multi… What is your favorite color? "Orange\nMagenta" "3\n5" 20595146
2 3657118 Multiple choice - multi… What is your favorite color? "Pink\nGreen" "1\n2" 19349250
3 3657118 Multiple choice - multi… What is your favorite color? "Blue" "7" 18334462
I've selected what I considered to be the most useful elements returned by the API to include in the result tibble. Each response contains both a text and numeric version of the participant's answer. Depending on the question type, one or both will be populated. The exact format of the values varies based on the question type.
The responses for a survey (or the responses for a subset of questions) can also be exported in a formatted document generated with Rmarkdown. Use the webq_report function to generate a basic report.