So… Poking around, I’m seeing three possible medical advances in treating this:
That seems promising.
My daughter and her fiancé live/work in Beaver County. At one point the only Covid cases in the county were at the nursing home.
It’s pretty hard to find a scatter plot without completely dependent variables that doesn’t have any outliers. That doesn’t change the trend line which shows a very real correlation. No “squinting” required.
homogeneous
End of hijack…
Decriptivism vs. Prescriptivism …
Good timing on the first bullet. JAMA Online today released multiple studies on this topic, plus an editorial summary of them:
Basically the corticosteroids are effective where there’s acute respiratory distress syndrome, but the appropriate dosage isn’t clear, and there’s still the issue of studies that show that the steroids also delayed recovery.
So progress, but not quite there yet.
Yes, and these drugs would only be used when people get very sick or after they recover from being very sick so that they don’t go through the reported crash. Other people would only receive the antivirals.
Yes. I think they might also be useful for patients with signs of stroke, renal and other cardiovascular issues. As you state, there is the risk of delaying recovery so they’re only being used when things get serious.
27,065,789 total cases
883,742 dead
19,165,962 recovered
In the US:
6,431,152 total cases
192,818 dead
3,707,000 recovered
Yesterday’s numbers for comparison:
Our new two-week lockdown here on Oahu is supposed to end this coming Thursday, the 10th. But no one thinks it will. Everyone expects a two-week extension. Our daily new-case numbers are still in the 200s and 300s every day, the vast majority on Oahu. The state’s lieutenant governor, who is an MD, says new cases need to be 150 or fewer per day to be assured of not overloading the hospital system.
I’ve tried to reproduce them, but with a trend line (and descriptive model) included.
FWIW I plugged the numbers into an app to calculate R and p (leaving out the EU as that seemed double dipping to include some member countries and the EU as a whole) and got a R roughly 0.75 with p 0.05.
There are 38 individual counties that have reported Q2 GDP on the source that @Trom provided. I would be nice to see the Deloitte person having included them all, rather than only a small subset of them. I suspect the association would be stronger, not weaker, if they did, but not including all the points always gives an eyebrow raise, especially when the apparent trend line is driven so much by just two European countries with both relatively very high death rates and exceptionally big contractions, and four Asian nations with low death rates and exceptionally small to no contractions.
I realize the point I am about to make may seem argumentative for its own sake, but I really do wonder - it seems to me that we might consider that the causation is as implied by having GDP change on the X-axis: fragile/vulnerable economies resulted in high death rates, while resilient economies were able to mount responses that kept death rates relatively low (exceptions being the resilient economies that by choice or poor leadership allowed greater spread to occur: Sweden and the U.S.)
I had the same reaction, so I just dropped 40 different countries into a graph. I was just coming back to the thread to make sure my numbers tied to what was there, I’ll post in a bit.
I’ve only been using Tableau for a little while, does this work?
Yes, and thank you for doing that.
Nice work! I downloaded your workbook and swapped the X and Y axes and reversed the Y axis to match the Deloitte plot. I also added a trendline to see if it still held true. Here is the sheet with those modifications.
https://public.tableau.com/profile/dmc7541#!/vizhome/GDPandCOVID_15994146814380/Sheet2
I had no idea you could download datasets from Tableau, is that always true? I’ve never thought to look there for data.
I added a trendline to mine. It’s funny how different they look with switched axes.
Just catching up on this thread. This is very good news, I’m surprised this study hasn’t gained more attention. Large unbiased sample, long period (4 months), multiple serological tests targeting different antibody isotypes and specificities… and it contradicts (and explains) earlier worrying studies that suggested humoral immune response may fade quickly.
This is good news for the likely efficacy of a vaccine, and for the probability of reinfection.
It’s up to the author what can be downloaded. Tableau Public is a bit different from the paid versions in that your data has to be embedded in the workbook (.twbx vs. a twb) or it won’t work. With the paid versions, that’s an option, but live connections are also fine.
Since you’re new, and since the majority of authors don’t typically lock down their Public work, you might want to find interesting graphs in the Gallery and click download to look at how they are done. The download button is on the bottom right, below the chart.
Thanks!