microsoft/Bing-COVID-19-Data

U.S. Death Data off by ~8k

Opened this issue · 3 comments

The US death count in the data is wildly under counted compared to BING search results and John Hopkins' tracker by about 8,000. Between BING and John Hopkins, for 5/11/2020 the deaths in the USA was about 82,000. The data here is showing 74,000. Is this data really from BING? If so, how did you obtain it? How is it updated?

74,000 for 5/11/2020 is off by about 1.5 weeks. The confirmed cases seems to be correct for 5/11. My guess is that the dates are just mislabeled for the death count by about a week and a half. Can you take a look?

The data set changed on 5/10, it no longer matches the Bing search results. All data seems to be restated back to January. I assume this is an error. Coincidently it changed when the Bing COVID page changed with new charts.

Additionally, I have been focused on analyzing US and NY. The NY numbers have remained consistent, but the US level aggregation has changed as described in my last comment. Seems something has changed with the aggregation to the US level.

Ok. What is the source of this data? Can I dig into this?
I don't know if you are aware, but Florida was just caught fudging the virus numbers down going back to May 5. I'm wondering if this has anything to do with shenanigans like that.