@EffectsFacts Yes. Where you have spread through a single population you get the Gompertz type behaviour. It’s follows logically if you understand just what Gompertz represents. IMO there’s a lot of significance in that observation, but will stick to what we’re discussing. All populations 1/n
@EffectsFacts 2/n are a mixture of sub populations to some extent. If if there is too much difference temporally or other ( E.g. US states...) then the combined overall curve is different, slower, not quite as neat, or a mess. So re scales of data knowing where that scale sweet spot ...
@EffectsFacts 3/n lies is valuable and tells u something about the place you’re looking at. National level works with UK, Italy etc... But national is obviously not the right scale size for analysis for the USA, for example and for other reasons you've pointed out ...
@EffectsFacts 4/n Re IL and PA in USA... data noisier. Cumulative curve fine. But yes slower. I wonder if local data would show more temporal spread between areas within these states. NY, for example, looks like it reached all corners v fast. In IL &PA the sub-waves, so to speak, ...
@EffectsFacts 5/n might be less aligned in time. If anyone knows a good source for data could be interesting to look.
...End.
...End.
@threadreaderapp
unroll
unroll
