Quantitative Editor, @FiveThirtyEight | PhD, @LSEGovernment | 
in 
, via 
| she/her
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Why is making a COVID-19 model so hard? @maggiekb1, @jazzmyth and I tried to unpack all the *many, many* sources of uncertainty that go into modeling deaths.
It feels like it should be somewhat straightforward...
https://fivethirtyeight.com/features/why-its-so-freaking-hard-to-make-a-good-covid-19-model/ …pic.twitter.com/EpUg9GXBhL
...but then "you discover that there isn’t a single number to plug into … anything. *Every* variable is dependent on a number of choices and knowledge gaps."pic.twitter.com/mvGWeUB4hp
Huge thanks to @WilbertVPanhuis, @maiamajumder, @drake_lab, Rae Wannier, @BillMiller_Epi, @irishmhw, @aetiology and @svscarpino for their knowledge, patience and kindness in this incredibly busy time!https://twitter.com/jazzmyth/status/1245027907489185792 …
Simply put, it's like a Monte Carlo simulation.
What about the rate of false negatives and false positives.
The admissions from Italy and France that their methods have been significantly undercounting COVID deaths should have gotten a lot more attention.
"Transmission rates" is something with a distribution - it could be that 2/3rds of us have an R<1, and small numbers of the very social mobile and irresponsible take the average up to 2.5 say. Patient 31 in South Korea, or that religious bell-end in India.
This is so beautiful and well-stated... Thanks much!
Great article. Thank you for the explanation!
Gee, it's almost like there is a differential equation in there somewhere.
Fantastic breakdown, now your ready for a bayesian analysis party next:)
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
