calpoly said:
Cal88 said:
calpoly said:
Cal88 said:
This is what I've said: "CA is actually on a similar trajectory than NYC, but has a later start to the epidemic than NYC". This statement, which you are nitpicking here, is correct, the slopes of the NY and CA charts having been very close in the last few days.
I am not just "taking just one segment", I am looking at the last few days of a data set that is less noisy because it is a plot with a 7-day moving average. This is clearly a very dynamic situation, so looking at the most recent segment of the graph is more relevant (and less noisy) than focusing on earlier stages.
This is NOT a linear scale so when you compare slopes you have to take that into consideration. A difference in slope on a non-linear scale is a BIG deal. You really don't know how to analyze data and I will leave it at that.
These slopes from NY and CA from the last 3 data points are virtually parallel, you can check that yourself by hand on a big monitor, as I have.
The last 3 data points actually cover the number of deaths in CA and NY over the last 10 days, with more weight placed on the middle of that period, and with useful smoothing through the 7-day cumulative tally. rendering those last 3 data points quite significant and less "noisy" as a representation of the current trend and a basis for future projections.
I can show you how to do this by hand if you'd like, we learn that in E36, sophomore year; you get a square and a ruler and slide the square parallel to the line segments over the ruler. You will see that the match between the two curves is very tight, confirming my assessment above from visual inspection.
And yeah we do know what a log scale is, and that small variations in slopes on a log scale result in bigger differences than on a regular scale, but any reasonably clear-minded scientist can definitely see that these two slopes have been trending very close together relative to all the other ones on this chart, you'd have to be an idiot (or a pretty bitter theoretical physicist) to keep denying this.
I think you're trying just a tad too hard to discredit my analytical skills here calpoly. Perhaps you could try to contribute some useful information to this thread and topic, instead of wasting your time and mine indulging in your petty ankle-biting.
"These slopes from NY and CA from the last 3 data points" This is the problem. You CANNOT throw away data because it does not support you hypothesis.
"I think you're trying just a tad too hard to discredit my analytical skills here calpoly. " No, you are really doing a fine job of displaying your ignorance with respect to data analysis without my help.
You're making the same mistake Nate Silver has made in his attempt to artificially inflate "red state" metrics by focusing on the higher percentage growth rates in states where the epidemic is in its noisier, embryonic stages, which was a highly misleading approach anda poor case made by Silver:
https://bearinsider.com/forums/2/topics/94392/replies/1739740The early stages of an epidemic tend to be a lot more noisy and less indicative of the general trend. For example, one infected patient zero member of a parish can end up infecting a dozen fellow parishioners, which would yield a 1,200% growth rate for that population, a rate that is not at all representative of future growth.
Silver tried to argue that states like Alaska or WV, which have had raw growth rates of 166% and 219%over a recent 3 day period were worse off than NY or IL, where the growth rate was 98% and 78%, but were the numbers were hundreds of times higher and the epidemic much further ahead.
Down the line in April we will get a more accurate picture of what the growth rate is in WV or Alaska and will be able to make a more valid comparison. The early stage numbers however are a lot less useful and can actually be misleading, as shown in Silver's misguided approach.
https://fivethirtyeight.com/features/the-coronavirus-isnt-just-a-blue-state-problem/The growth rates eventually stabilize and become less noisy as time progresses and the number of the infected population grows, reflecting the real nature of the epidemic's "terrain" in each state.
That's why looking more closely at more recent data from the last 10 days, which I did above, is not the exercise in futility that you infer, but actually the best snapshot and indicator of future trend, considering the epidemic was still in its very early stages in early March (especially in CA).