Hong Kong University has a “real-time” with a plot of R effective measured every day!

That’s impressive.

Plotted on the same time axis as the case plots below so you can match them up.

https://covid19.sph.hku.hk/dashboard  https://covid19.sph.hku.hk/dashboard 

I take back the axes match up (I didn’t look closely enough).

The one thing this does point up (and we’ve seen this in Seattle too) is the estimates (from smartphone data?) of interactions which feeds into the estimate for R seems a bit high.

Compare cases dropping but R ~ 1

Think about how R (the number of infections on average passed on by one infectee) is calculated:

R = D * O *T * S

D = Duration of infectivity (time units)
O = Opportunities for infection per time unit
T = Transmission probability
S = Susceptibility of the population

If you are interested in a intelligent general readers view of epidemology, or a Theory of Happenings as Ross had it, I strongly recommend @AdamJKucharski book The Rules of Contagion. It covers a lot more than just disease (fashions, fads, bubbles, etc).

http://www.kucharski.io/books/ 

So you can put an estimate on how much change in solcial interaction you need to get R effective below 1 (for the outbreak to fade away exponentially).

For R0 = 2.5 you need a reduction of more than 60%

And the advice to King Co WA the “outbreak stopper” was 75% reduction.

Another example of how you can see the R = DOTS helping you to think about how effective can be changed is NPIs (social distancing and staying home) reduces Opportunities to pass the disease.

If you can reduce O by 1 - 1/R0 then you can get R = 1 https://twitter.com/kevinpurcell/status/1247231506038153220