I made another plot that includes seasonal baseline, but haven't been updating it. The COVID peaks were abnormally large. https://pasteboard.co/7BVwLNxiwyhR.png
Your bar graph also shows the rise in 2020 after a five year decline.
I made another plot that includes seasonal baseline, but haven't been updating it. The COVID peaks were abnormally large. https://pasteboard.co/7BVwLNxiwyhR.png
Your bar graph also shows the rise in 2020 after a five year decline.
yes, but to a level that was perfectly normal over the last 5 or so years and after an abnormally low year in a series that exhibits strong mean reversion from outliers because of the effects on vulnerable cohort size.
the simple fact is this:
if you had swapped 2020 for 2018, 17, or 16, no one would have noticed.
2020 was absolutely average in comparison to them.
that's a massive win in a time when many place (like the US) saw a 25% excess death figure.
you seem determined to find a crisis where none occured.
I went into this analysis thinking COVID panic was overhyped and that is still my view. But there is a significant mortality signal that shouldn't be ignored or obfuscated.
Adjusting for population growth would only make sense if the demographic that grew was not young. I've been hearing a lot about immigration in Sweden, and immigrants tend to be younger and have more kids.
2020 is clearly an outlier on the all-cause mortality graph that I made, with 3x as many deaths in the COVID peak as the 2018 flu peak. Then there was another unusually large peak going into 2021. There is no summer peak.
I did visit Twitter earlier in the year, and it was pretty clear that a number of voices were working hard to show graphs that made the 2020 rise in Sweden seem like noise. That is not science.
I made another plot that includes seasonal baseline, but haven't been updating it. The COVID peaks were abnormally large. https://pasteboard.co/7BVwLNxiwyhR.png
Your bar graph also shows the rise in 2020 after a five year decline.
yes, but to a level that was perfectly normal over the last 5 or so years and after an abnormally low year in a series that exhibits strong mean reversion from outliers because of the effects on vulnerable cohort size.
the simple fact is this:
if you had swapped 2020 for 2018, 17, or 16, no one would have noticed.
2020 was absolutely average in comparison to them.
that's a massive win in a time when many place (like the US) saw a 25% excess death figure.
you seem determined to find a crisis where none occured.
So I updated the plot, and the conclusion that COVID-19 caused a lot of excess death in Sweden is unchanged. https://pasteboard.co/4uu4G3z7XXs6.png
I went into this analysis thinking COVID panic was overhyped and that is still my view. But there is a significant mortality signal that shouldn't be ignored or obfuscated.
Two more weeks: https://pasteboard.co/kw64pqLEZIUM.png
you're not even adjusting that for population growth and that kind of overly smoothed baselining ALWAYS shows misleading peaks.
you're really mishandling the data here and mistaking shifted seasonalities for actual excess deaths.
modeling as you are means that a summer (vs a winter) pandemic will always look huge because it's vs a traditionally low baseline.
what you really need to know is AUC.
try following @haraldofworld
he can show you how to do this for real.
Adjusting for population growth would only make sense if the demographic that grew was not young. I've been hearing a lot about immigration in Sweden, and immigrants tend to be younger and have more kids.
2020 is clearly an outlier on the all-cause mortality graph that I made, with 3x as many deaths in the COVID peak as the 2018 flu peak. Then there was another unusually large peak going into 2021. There is no summer peak.
I did visit Twitter earlier in the year, and it was pretty clear that a number of voices were working hard to show graphs that made the 2020 rise in Sweden seem like noise. That is not science.