General Discussion
Related: Editorials & Other Articles, Issue Forums, Alliance Forums, Region ForumsJohns Hopkins Covid-19 Graph Takes a Disturbing Turn
5 Day Moving Average for confirmed new cases-Happy Easter
scroll down to first graph
https://coronavirus.jhu.edu/data/new-cases
entry error? seems unlikely the 5 day average doubled in one day.
Ms. Toad
(34,086 posts)likely a data entry error. But one would think they would do a direct pull from the database supporting their dasboard . . . (which has the correct number (or at least a number within the range the other sites are reporting.)
Ms. Toad
(34,086 posts)If you look at the "actual data" on 4/12, you will see that they report 127,306 new cases. The new cases are running in the 20 thousands, not the 120 thousands. On worldometers - a site I switched to because of stupid stuff like this (as well as less regular reporting), the number of new cases on the 12th was 26,641. On the Johns Hopkins Dashboard, it was 28,9XX.
I always do a reality check on the numbers. For the 5-day average to jump as high as the graph indicated, there would have to have been about 5x as many NEW cases on the 12th as in any preceding day. The graph for new cases indicates that is correct - so then I do a check against other sources - none of which support 127,306 new cases on the 12th.
Thus the data is wrong and, as they say - garbage in, garbage out.
FreeState
(10,575 posts)27,000ish is much more likely than 127,000ish.
LeftInTX
(25,515 posts)It doesn't seem right. The actual percentage of new cases is going down (Not the actual number of new cases), so you wouldn't see a crazy slope like that.
For instance:
City A on Thursday you have 100 cases. On Friday you have 150. That's a 50% increase in one day The slope would be 1/2x
Three weeks later:
1000 on Thursday and 100 on Friday is a 10% increase the slope would be 1/10x
duforsure
(11,885 posts)Like nursing home deaths ,and from other groups?
Just a typo. If you look today, the new case # for 4/12 is now 25,306.
(A stray additional 1 (00,000) cases, and the hundreds was actually 3 (00), not 5 (00))