State by state estimated R_t

1,971 Views | 9 Replies | Last: 5 yr ago by Keegan99
Keegan99
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
9:20p, 4/18/20
AG
https://rt.live/

Digging in to this now. Sensible Bayesian methodology.

At first first blush, it's interesting how New York and Louisiana were both on fire, so to speak, but have now plummeted. Is that due to extremely vigilant behavior or the virus simply running its course?
Reply Quote
7
Bobcat06
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Keegan99 • 9:35p, 4/18/20
AG
My guess is behavioral changes.

NY is at 1.2% infected and LA is at 0.5% infected. I would expect a much higher infection percentage for growth to slow due to saturation.
Reply Quote
4
FHKChE07
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Keegan99 • 9:45p, 4/18/20
AG
This is just a crapshoot as the first round of anti body testing in california was seeing rates of 50-80 times as much as what is actually measure. You can't estimate R values if you have no idea infection rates.
Reply Quote
2
Keegan99
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to FHKChE07 • 9:51p, 4/18/20
AG
You can infer that reported cases are reflective of actual cases in some proportion.

If reported cases are declining, all cases are declining.
Reply Quote
7
DadHammer
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Keegan99 • 9:55p, 4/18/20
AG
I am guessing it's running it's course. Plus social distancing and smarter hygiene.
Reply Quote
4
1 edit
Keegan99
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Bobcat06 • 9:58p, 4/18/20
AG
You assume that testing is detecting all (or most) cases. NYs higher percentage may be a result of higher test availability.
Reply Quote
1
FHKChE07
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Keegan99 • 10:03p, 4/18/20
AG
To an extent but it seems like when there was no testing, everyone was clammoring for testing and that ramped up exponentially, but after that people stopped caring to be tested because it doesn't change anything and so now there aren't as many people getting tested. Also, the initial ramp of testing is catching the backlog of all the people who already had it which was 2 weeks worth of people. Furthermore, the totally random way that they are reporting the data on the public testing which is being reported back in batches back to the governments with no respect to when the tests were actually testing has got to mess with all of this.

I just think that people are trying to extract anything meaningful out of a large pile of severely crappy data that is varied from state to state and generalize it when there is a lot of assumptions being made. Especially when in a lot of places we are talking about statistically insignificant data. 3000 positive cases over 6 weeks in a MSA of like 6 million people? When the real number of people that have likely been infected is probably close to 2 orders of magnitude more than that?
Reply Quote
2
Bobcat06
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Keegan99 • 10:24p, 4/18/20
AG
Keegan99 said:
You assume that testing is detecting all (or most) cases. NYs higher percentage may be a result of higher test availability.
I assume that the number of positive tests is roughly proportional to the number of actual cases.

Louisana went from a peak rt of 2.7 three weeks ago to rt of 0.8 currently. If behavior is unchanged, a drop of that scale would require 70% of the population to be infected (vs current 0.5% reported infection rate).

I believe that there are testing issues and the number of actual cases may be 2 or 3 times higher than reported cases. I don't believe the number of cases is 140 times higher than reported.
Reply Quote
2
Bobcat06
User Profile Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Bobcat06 • 10:27p, 4/18/20
AG
Also worth noting, that in the last 3 weeks the CDC reversed it's recommendation on masks. It went from being taboo to nearly universal.

I suspect that's the biggest impact on rt goin below 1.
Reply Quote
3
FHKChE07
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Bobcat06 • 10:28p, 4/18/20
AG
Bobcat06 said:



I believe that there are testing issues and the number of actual cases may be 2 or 3 times higher than reported cases. I don't believe the number of cases is 140 times higher than reported.

https://thehill.com/homenews/news/493391-california-antibody-testing-suggests-coronavirus-may-be-far-more-widespread
Reply Quote
Keegan99
User Profile
Private Message
Ignore User Stop Ignoring
How long do you want to ignore this user?
24 hours One week Permanently Cancel
In reply to Bobcat06 • 10:57p, 4/18/20
AG
And I don't believe that's a sound assumption since testing availability is not geographically uniform.
Reply Quote
1
Refresh
Page 1 of 1
 
×
subscribe Verify your student status
See Subscription Benefits
Trial only available to users who have never subscribed or participated in a previous trial.