Je kan informatie over je locatie aan je Tweets toevoegen, bijvoorbeeld je stad of exacte locatie, via het web en applicaties van derden. Je kan altijd de locatiegeschiedenis van je Tweets verwijderen. Meer informatie
Since the whole world has gotten into modeling, this is a great article @TheAtlantic laying out why we should never think of these as trying to exactly predict the future: https://www.theatlantic.com/technology/archive/2020/04/coronavirus-models-arent-supposed-be-right/609271/ …
1/n
They're useful tools to compare potential scenarios, but since we're dealing with exponentials, tiny changes can have huge effects on outcomes. Here's one example where I start with 1k cases and project 2 weeks out based on different doubling times - e.g. diff 1.5 vs 2.0! 2/npic.twitter.com/WDZM35pD3d
So small differences in (assumed) doubling times will have massive effects on projected outcome numbers. This is very important for #COVID19 since most (undersampled) places start out at ~1.5-2.0, then 2.5, 3.5, etc. before averaging at 7+ days prior to plateau and decline.
3/n
... and here I'm just doing a straight up regression with a single variable (doubling time) - real models are much more complex and have many more variables, most of which are uncertain at this time. YET, e.g., @MRC_Outbreak have managed to provide many very useful models!
4/n
Note - I'm in no way a modeler, so take all of what I just said with that in mind - including if you double-check my calculations above... 
5/n
It would be great if we can have another table showing how much impact different measures such as extend of isolation have on doubling time.
Twitter is mogelijk overbelast of ondervindt een tijdelijke onderbreking. Probeer het opnieuw of bekijk de Twitter-status voor meer informatie.
