Urban areas that seem to qualify: 1. [UK has only 3](https://en.wikipedia.org/wiki/List_of_urban_areas_in_the_United_Kingdom) 1. [The EU has 20](https://en.wikipedia.org/wiki/List_of_urban_areas_in_the_European_Union) (if I counted properly) 1. [The USA has 21](https://en.wikipedia.org/wiki/List_of_United_States_urban_areas) (of which _San Juan, Puerto Rico_ is one) So a total of 44. I could well be missing some because of the way those urban areas are defined on those lists _vs._ how they are defined on the individual Wikipedia pages. _Edit:_ Fixed ...

@Jgalt I wouldn't even worry about whether or not the samples are present. That fridge door seal seems to be at least 3rd level containment (after plastic vial and the box). I doubt that "broken" seal made any substantial difference.

Also, the lack of a positive pressure personnel suit suggests whatever is in those vials is not so dangerous.

@Jgalt More of the same from Reuters: https://www.reuters.com/article/us-health-cor… :

General Mark Milley, chairman of the Joint Chiefs of Staff, said on Tuesday that U.S. intelligence indicates that the coronavirus likely occurred naturally, as opposed to being created in a laboratory in China, but there is no certainty either way.

[...]

The broad scientific consensus holds that SARS-CoV-2, the virus’ official name, originated in bats.

New weekly report (their emphasis):

In the week ending April 4, the advance figure for seasonally adjusted initial claims was 6,606,000 a [...]

Does it make sense the use the seasonally adjusted values for computing predictions?

My guess is yes, since the value we are predicting is seasonally adjusted IIRC, although I have been using the unadjusted values in my toy model: https://www.getguesstimate.com/models/15723

Feedback welcome!

This started a few days ago already: [Australia to pursue coronavirus investigation at World Health Assembly](https://www.reuters.com/article/us-health-coronavirus-australia-china/australia-to-pursue-coronavirus-investigation-at-world-health-assembly-idUSKCN2251G7) Even though Australia is not on the resolution criteria list: > Australia sits on the executive board of the assembly, which determines WHO policies and appoints the director-general. The assembly is due to meet on May 17. > [...] > Australia has overcome the worst of its coronavirus outbre...
Interesting news from [The Guardian Live Feed](https://www.theguardian.com/world/live/2020/may/04/coronavirus-live-news-mike-pompeo-pushes-virus-lab-theory-as-brazil-passes-100000-cases?page=with:block-5eb036878f0826dcf8dc2d12#block-5eb036878f0826dcf8dc2d12): > A French hospital has retested old samples from pneumonia patients and discovered that it treated a man who had Covid-19 as early as 27 December > [...] > Cohen said the patient had survived and that an investigation to trace the first contamination has been carried out. He said: >> He was sic...

Reduced uncertainty as we now have a new data point with latest weekly figures.

My relatively naïve model is here: https://www.getguesstimate.com/models/15723

I've been using the unadjusted weekly numbers, which might not be wise, since the final result is seasonally adjusted, but I couldn't be bothered to factor those changes in :D

Some interesting bits and commentary from _The Economist_ from yesterday ([Covid-19 is spreading to America’s South with unnerving speed](https://www.economist.com/united-states/2020/04/21/covid-19-is-spreading-to-americas-south-with-unnerving-speed)): > In the ten days after the first confirmed case in the Big Easy, the number in Louisiana grew faster than anywhere else in the world, says Gary Wagner of the University of Louisiana at Lafayette. In mid-April, the New Orleans area had the highest per-person death rate from covid-19 of any American city, ...

@AABoyles The COVID Tracking Project launched out of The Atlantic is another source.

If the various sources all corroborate that should work.

Current daily deaths averaged over 6 days is 55-106 (.25-.75 confidence based of std. dev.) That gives me 1045-2014 deaths over the next 19 days, with 1,349 deaths declared to date.

I've adjusted that prediction down ever so slightly, but between the slowing death rate, and more tests, I'm assuming things will stay relatively constant for a while.

It sounds like designing an appropriate [observational study](https://en.wikipedia.org/wiki/Observational_study) could work. This would be based on either existing or future interventions. State-wide is a good starting point, but where county or city interventions exist those could also be evaluated. _Edit:_ Since you asked for specific suggestions, and I failed to provide any, here is one that is not super well thought out, but is a start: > What will be the median difference of Covid-19 cases per capita in 2020 between states that imposed "stay-at-...
From the BCC: [Europe 'wary of confronting China over deaths'](https://www.bbc.com/news/world-asia-china-52404612) > China is continuing to under-report the true levels of deaths from Covid-19, national security officials in London and Washington believe. > But European officials are wary about directly challenging China over the figures, sources told the BBC. > [...] > Governments in Europe have been wary of engaging in a diplomatic row with China at a sensitive time. One UK official, who like others declined to speak on the record, said there was a...

From the UK Gov page:

Data from around 5 days ago can be considered complete. Data for recent days are constantly being revised as more information becomes available.

This is clear from looking at the data, and since the question seems to be clear, I take it that the result is as visible on the day, not after stabilisation adjustments.