Part 1 Intro
It's not about the poster, 'sarah hucklebee', it's about interpreting the CDC data as reported. We detected wobbly data, so went to the source. But first, let's recap.
Part 2 Late Night Math Errors Proliferate
Your guessing was unfairly close, but we find for the following: As the information is not static, and goes through data point revisions and error scrubs, it's a moving target! Consequently, the CDC numbers are 'as accurate' as can be when reported, yet when updated, even this snapshot post won't wash!
The initial data presented was a summation, so we did custom CDC searches at the [expert] level and found the raw source data at difficulty factor 6. Every precious comment on this mortality topic /slide, (except for greeneyes), made seemingly 'minor' mathematical errors which massively skewed the results.
The #1 analysis -since last night- presents excellent estimations, but had a precision error which jacked up the EOY estimated deaths. Not to belabor, but it's not 0.006, but actually 0.00556317701610132, which gives us 2870883 total, 7843/day.
That difference (between .006 and .005563) is 1247 lives.
Every life is important.
A further comment incorrectly guesses a newly introduced factor of .264, which is too low- the actual factor is .2677596.
That's a difference of 10,734 lives.
Every life is important.
Part 3 Towards A Data Approach
Source of CDC Monthly Mortality numbers for 2019 and part of 2020 can be found here:
https://www.cdc.gov/nchs/nvss/vsrr/provisional-tables.htm
With a "12-month ending" total we calculate an average standard deviation BETWEEN each month of .06 ( 6% ), yet see too much variance in the source data month-to-month (MTM) before the 'rollup' into greeneyes survey.
Examples for this include:
- A ONE (1) month swing of 19% in 2019 Feb. - Mar. 2019
- High MTM variance in Feb. 2019: 9.69%, and Apr. 2019 -7.84
- High MTM variance in Mar. 2020: 10.2%
- VERY LOW variance in Nov. 2019 of .42%.
From this we may conclude that the 'total years' number from greeneyes rollup report has smoother /more accurate data unencumbered by unreported /lagging updates, and that there is, in fact, no "EOY" JUMP in numbers.
Just think of it, greeneyes, you just saved somewhere between 1,247 to 10,734 lives or more in 2020.
Bravo!
