search

starschema / COVID-19-data

Unpivoted and cleaned data sets on the COVID-19 pandemic
https://starschema.github.io/COVID-19-data
BSD 3-Clause "New" or "Revised" License
85 stars 19 forks source link

Kings County New York in JHU_COVID_19 reporting 0 cases #241

Closed psarnow closed 3 years ago

psarnow commented 4 years ago

Describe the bug There may be more that 'seem' incorrect, but King's is reporting many cases as indicated by JHU's own dashboard. Here, we have values of 0 cases, and I can't tell why.

To Reproduce

SELECT * 
FROM jhu_covid_19
WHERE FIPS = '36047';

Expected behavior Expecting our target table available here to match what JHU displays. https://coronavirus.jhu.edu/us-map

Screenshots image

Additional context I've been searching for 'exceptions' from JHU, but I can find anything pertaining to this.

tfoldi commented 4 years ago

We are using the daily data from JHU's repository: https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_daily_reports/05-05-2020.csv

This file does not have Kings county in New York

psarnow commented 4 years ago

Understood. Problematic that King's county and the other New York City boroughs have rows in the table with case counts of zero though, that is terribly misleading.

psarnow commented 4 years ago

Are you aware if any of the tables provided by Starschema here have New York City county level case data? I'm at a loss so far.

tfoldi commented 4 years ago

The most granular is NYC_HEALTH_TESTS that has zip code level info too

tfoldi commented 4 years ago

It gets data from https://github.com/nychealth/coronavirus-data

psarnow commented 4 years ago

awesome. thanks!

psarnow commented 4 years ago

Understood. Problematic that King's county and the other New York City boroughs have rows in the table with case counts of zero though, that is terribly misleading.

May I suggest instead of filling missing FIPS from the daily data in JHU's repository, leaving it out. Not reported is much different than reporting a value of 0. If the file does not have Kings county in New York, then there should not be a row in the resulting table.

tfoldi commented 4 years ago

sounds good, we will remove those old legacy counties with 0 cases

psarnow commented 4 years ago

Great! To be clear, don't remove rows from JHU's repo (even if they have zero cases). Just don't create new rows that are not in the JHU repo. This will help a lot. Thank you.