Skip to content

Conversation

@jingjtang
Copy link
Contributor

  • Pull data from the email attachment
  • The pulling step is very time consuming especially when we have much more dates to consider, so now:
    • store the intermediate file (aggregated by zip with number of total tests, number of positive tests) as cache
    • only update new data when run next time
  • ~1.97% invalid tests, Now treat them as negative tests
  • Geographical poolling is added since the data volume is now enough. The minimum sample size is 50. If there is no data report for a state/hrr/msa/county, it is filtered out because of this.
  • The data start from 5/9 but has only extremely low volume. So, temporarily set the export_start_date to be 5/26.
  • As for the time stamp:
    • Use test date as time stamp for normal cases
    • If storage date is earlier than test date which is abnormal, remove those cases. (0.10%)
    • If storage date - test date > 90 days. (0.01%)
  • Report data through today (GMT) minus 5 days due to backfill
  • Do we want to report county level data? (did not report county level for flu tests due to the data volume problem, there is still the same problem for COVID Test)

Copy link
Contributor

@krivard krivard left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good! This will update from 5/26 to (today minus 5 days) each day it is run, right? I believe that's what we want.

Other todos:

@krivard krivard requested a review from eujing July 16, 2020 18:07
@jingjtang jingjtang requested a review from eujing July 16, 2020 21:50
Copy link
Contributor

@eujing eujing left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me!

@krivard krivard merged commit fae9e1d into main Jul 25, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants