6a24d4fa23
chore: Make sure 'hispanic' column only gets applied when using SIS race information
2 years ago
a15b01a3e1
fix: instead of looking for 'asian' at the start of a word, look for it
...
after a word boundary. This means it still doesn't get confused with
caucasian and it's more flexible whan asian appears inside other text
such as 'Caucasian and Asian and Black'
2 years ago
8a0ba0dbea
chore: various fixes for race and gender categorization during cleaning.
...
Also add tests for race and gender categorization
2 years ago
2ef24caf70
Lower threshold for the number of valid student responses from 17 to 11
2 years ago
f27a590c5a
Convert gender and race text into qualtrics codes during cleaning. Abide by 'prefer not to disclose' for self reported race. Give priority to self reported data but use SIS information as backup
2 years ago
a3f9e46414
chore: when searching for dese id, split up pattern so that to be more explicit about the order in which to search out the columns that might have the dese ID we're looking for.
2 years ago
b2fdbe5756
feat: We no longer trust the progress number that gets exported from qualtrics. Instead during the cleaning progress, perform a manual count of the number of responses to filter out rows that don't meet the minimum threshold.
2 years ago
ef44c41965
feat: add special education disaggregation
2 years ago
2fd56047d4
Add disaggregation by ELL
2 years ago
2ac30bb107
feat: Add income table to the database. Add seeder for income. Add a reference to income from survey item response. Update the loader to import income data from the survey response csv. Refactor analyze controller to extract presenter. Add corresponding specs. Add income graph to analyze page
2 years ago
0f23053294
It's possible for admin data likert score values to be above 5. If that happens, we
...
cap the likert score at 5. This was happening already at the scraper
level but it's also now being done by the admin data loader for safety.
Also make sure to just update admin data instead of deleting and
reloading all values each load. Add tests to confirm this behavior
3 years ago
e058c523b6
Missing progress or duration information does not result in a row removed in the cleaning process
3 years ago
a71ebbc4e4
Add Overall Response Rate
3 years ago
dbfc9d1d3a
Add automated data cleaning. Modify SurveyItemValues class to use regex
...
instead of hard coded values. Produce a clean csv and a csv with all
the removed values and columns with reason for removal. Add script for
running cleaning for each project
3 years ago
Nelson Jovel