search

dkobak / excess-mortality

Excess mortality during COVID-19 pandemic
GNU General Public License v3.0
331 stars 45 forks source link

No data for some countries #4

Closed vovodroid closed 3 years ago

vovodroid commented 3 years ago

There is no Excess per 100k and Excess as % of annual baseline columns for some countries, for instance Finland, Norway and more.

dkobak commented 3 years ago

For these countries the excess mortality is not significantly different from zero. We decided not to provide these normalized measures in this case, as they can be misleading (e.g. look misleadingly large even though in fact not significantly different from zero.)

vovodroid commented 3 years ago

look misleadingly large even though in fact not significantly different from zero

Why? How 0.5% would be misleading?

dkobak commented 3 years ago

You are probably right that these two columns (per 100k and as %) won't be misleading. What can be misleading is the "undercount ratio". For example, in Hong Kong it comes around 20, because the excess mortality is ~2000 (not significantly different from 0 though) but the official number of covid deaths is ~100.

I could fill it in the per100k and the % columns and only leave nans in the undercount column. Or fill everything in and hope nobody will be misled.

What do you think? CC @akarlinsky.

vovodroid commented 3 years ago

So what is decision?

dkobak commented 3 years ago

Thanks for the reminder. I have added to the CSV file the excess per 100k and the excess as a % of annual for countries where excess is not significantly different from zero.

vovodroid commented 3 years ago

Great, much better know!

By a way, are you preparing for second year circle? It will be here soon.