[FRIAM] What Is the Real Coronavirus Toll in Each State? - The New York Times

[uǝlƃ ☣]

I don't think so. Doubling time (under non-ideal lockdown conditions) conflates density with the other measures. If a sparsely populated area has the same doubling time as a densely populated area, the doubling time graph might gloss over how *badly* the sparsely populated area is doing in controlling the spread. So, in order to really grok the doubling time graphs, you have to have a feel for the relative densities. Feel free to correct my faulty thinking.

I think the deltas (as in the graphs I've posted) are a good compromise. They're still in the same units (# of people) and show larger deltas for larger populations. However, I *would* like to divide out area (e.g. square meters) of whatever region's being plotted. I think Δcases/m^2 would be interesting.

On 5/6/20 2:49 PM, thompnickson2 at gmail.com wrote:
> Doesn’t doubling time handle that problem? 

> On 5/6/20 12:59 PM, Steven A Smith wrote:
>> 
>> This just underscores how hard it is to make sense out of absolute numbers, or more to the point, numerators without denominators.    At least some of the charts of absolute numbers (as long as they are not renormalized from situation to situation) provide a visual estimation of "slope".

-- 
☣ uǝlƃ