Saturday, April 25, 2020

COVID-19 Deaths as compared to other causes of Death (US Only) Week 8 Week update

Below is weekly and annual comparison data from 4/17/2020 - 4/23/2020 @ 9:47 AM PST 4/25/2020

Note: Averages have a *; Comparative Analysis and Pneumonia are highlighted in blue; COVID-19 is highlighted in Yellow


Weekly



Annual




Analysis, Benchmarks, and Factor Analysis


As of the end of Week 8:

  • COVID is around 9.4 times more deadly than the High Flu peak weekly deaths benchmark 
  • COVID is the most prevalent form of death in the US based on averaged benchmarks
  • COVID deaths have remained roughly even since last week (factor 1.02)
  • COVID deaths exceed the number of average annual deaths for Firearms, Suicide, and Pneumonia.
Factor Analysis - below is a 'factor' analysis for COVID annual deaths versus benchmarks.  

Numbers which COVID does not exceed will be presented as X%.  This denotes that the COVID mortality rate equals X% of the higher number.

Numbers that COVID does exceed will be presented as N.  This denotes that the COVID mortality rate is N times higher than the benchmark number.  A one (1) means that the numbers are equal, a two (2) that COVID is twice as deadly on an annual basis, etc.

All numbers are rounded to the nearest decimal.


Factor Analysis





Other info


Charts

I have received some feedback that visualization may help folks understand numbers better.  I am testing a few chart ideas. 

The first is a weekly death rate growth chart.  I am trying to provide information not being presented in other locations and I have noticed almost all charts online are cumulative.  The chart below shows rough week-to-week data. 

The other chart below is a visualization of the Annual Table above without Cancer and Heart Attacks.  This is due to the 'skew' those numbers present because they are so much higher than the others.  They are mainly on the Annual Chart to show the two major causes of death in the US as comparison and to help give context to weekly numbers.  

As always I welcome additional feedback to make the information presented here more relevant and easy to understand. 






State Variability

State variability is becoming more of a factor in reporting totals.  As reported in Week 6 there is a great deal of variability depending on what State you are in.  As States open up at different rates I think it is important to show changes to this variability over time.  I have had difficulty tracking this since WorldMeter does not break State data down over time, just a total for the day.  However, I believe I have a way to track week to week now.  

Starting this week I will be including a 'Top 10' list to show where deaths from COVID are most prevalent and to show changes over time.  

There are still limitations in this data.  In particular, New York/Jersey are really centered around the New York City (NYC) metro area.  So while their numbers are separated, they can be seen as the same 'cluster.'  Also, both counts can't really get the NYC cluster out from the rest of the State.  If I can figure out how to do it, a better way of displaying this would be to show NYC, then everyone else.  I will continue to work on that as I move forward. 

Finally, it will be easy to make correlations from the data presented below.  In particular, 'colder,' States are bearing the brunt of this as are higher population States.  I would urge caution in both of those assessments.  While there might be some relation in both cases, experience tells me there are a number of other factors that are at play.  


Top 10 States by COVID Fatalities




Note: Change in Rank for this week is from Week 6.  Future Tables will be from the previous week.  Red highlights mean the State has moved higher, Green highlights mean the State has moved lower. 

Corrections 

Two major corrections to previous data for this week.  One, I got 'off-cycle' a bit a couple of weeks back and added a extra day into Week 5.  It didn't have a major impact on any factor numbers or placements; however, I wanted to be transparent. It has been corrected moving forward. 

Second, WorldMeter was able to back-fill a lot of the data from New York into their daily reporting. As such I was able integrate about half the numbers into previous weeks data.  You will see this reflected in the 4/25 reconcile.  One side note, while I don't revise weekly placements, Week 6 saw COVID move from #3 to #2.  Overall the Raw Annual number is much closer to the stated number without adjustments. 

Finally, just a note that I am doing this as a side project in my free time.  Errors and mistakes (particularly spelling and grammar mistakes) will occur but I will be as transparent as possible when I catch them or am alerted to them.  

Also remember this is a on-going event.  Changes should be expected, even large number adjustments.  As I have said before, treat all numbers as preliminary and with an error range. 


Next Update: May 2nd by 6 PM Pacific

Saturday, April 18, 2020

COVID-19 Deaths as compared to other causes of Death (US Only) Week 7 Week update

Below is weekly and annual comparison data from 4/11/2020 - 4/17/2020 @ 12:40 PM PST 4/18/2020
Note: Averages have a *; Comparative Analysis and Pneumonia are highlighted in blue; COVID-19 is highlighted in Yellow


Weekly



Annual



Analysis, Benchmarks, and Factor Analysis

As of the end of Week 7:
  • COVID is around 9 times more deadly than the High Flu peak weekly deaths benchmark 
  • COVID is the most prevalent form of death in the US based on averaged benchmarks
  • COVID deaths have increased by a factor of 1.3
  • COVID deaths exceed the number of annual deaths for the benchmarked Low-Moderate and Moderate Flu and average annual deaths due to Car Accidents. 
Factor Analysis - below is a 'factor' analysis for COVID annual deaths versus benchmarks.  

Numbers which COVID does not exceed will be presented as X%.  This denotes that the COVID mortality rate equals X% of the higher number.

Numbers that COVID does exceed will be presented as N.  This denotes that the COVID mortality rate is N times higher than the benchmark number.  A one (1) means that the numbers are equal, a two (2) that COVID is twice as deadly on an annual basis, etc.

All numbers are rounded to the nearest decimal.

Factor Analysis





Other info

Annual Increase Corrections and Transparency 

New York and several other jurisdictions have started revising previous numbers of reported deaths.  My data source is not able to dis-aggregate by date so just added them to 'one' date this week.  That would unfairly skew this weeks numbers.  The site stated that future adjustments would occur like this but be called out.

As such additional reported deaths that occur in large 'batches' will be added to the annual number only to prevent weekly numbers from being skewed.

For transparency I will start include a running list of adjustments made to the annual total and the rational. If I am unable to match the stated annual number at the time of post (usually due to one of these mass adds or older updates) I will make a 'reconcile' to match and notate in the chart.

Please note that numbers will continue to be adjusted as corrections come in.  Unless it is the current week of, all adjustments will be made to the annual total only.  That said, I may start making the underlying data tables available at a later date.

As a result, please treat the 'weekly' totals as preliminary and the 'annual' totals as the current estimate.

Corrections Chart



Saturday, April 11, 2020

COVID-19 Deaths as compared to other causes of Death (US Only) Week 6 Week update

Below is weekly and annual comparison data from 4/4/2020 - 4/10/2020 @ 1:54 PM PST 4/11/2020
Note: Averages have a *; Comparative Analysis and Pneumonia are highlighted in blue; COVID-19 is highlighted in Yellow


Weekly



Annual



Analysis, Benchmarks, and Factor Analysis

As of the end of Week 5:
  • COVID is around 7.1 times more deadly than the High Flu peak weekly deaths benchmark 
  • COVID is the 3rd most prevalent form of death in the US based on averaged benchmarks
  • COVID deaths have increased by around a factor of 2 since Week 5
Factor Analysis - below is a 'factor' analysis for COVID annual deaths versus benchmarks.  

Numbers which COVID does not exceed will be presented as X%.  This denotes that the COVID mortality rate equals X% of the higher number.

Numbers that COVID does exceed will be presented as N.  This denotes that the COVID mortality rate is N times higher than the benchmark number.  A one (1) means that the numbers are equal, a two (2) that COVID is twice as deadly on an annual basis, etc.

All numbers are rounded to the nearest decimal.

Factor Analysis



Other info

Variability

Just a quick note on variability.  The numbers presented here are for the nation as a whole; however, diving into specific state by state data will show vastly different situations.  For example, for folks in New York, COVID is most likely the number one cause of death currently; however, for someone in Wyoming it may not even been registering on the list.

So far, 66% of deaths are occurring in five states (in order):

  • New York (7,844 | 42%)
  • New Jersey (1,932 | 10%)
  • Michigan (1,281 | 7%)
  • Louisiana (755 | 4%)
  • Massachusetts (599 | 3%)

Over time New York's share will naturally lower as other states reach their peaks.  In particular, Louisiana, Georgia, and Florida have seen large increases.  I expect the numbers above to flatten or recede in the coming weeks; however, the make-up of where those numbers are coming will be changing and a overall lowering may be some time off.  This may become a regular part of the site if I believe that the variability is an important metric to report on.

To be clear though, this is site is for comparison only and as such will not go into projections.

Future Updates

Now that the 'peak' has been reached for COVID in the US I will be converting this to a weekly updated site.  I believe there is still value in producing this; however, the need to show status updates on a bi-weekly basis has receded.  If, for whatever reason, COVID continues to spike in the US and/or spikes in the future, I will bring the bi-weekly updates back.

The next update will be by 6 pm PST on 4/18/2020.

Wednesday, April 8, 2020

COVID-19 Deaths as compared to other causes of Death (US Only) Week 6 mid-Week update

Below is weekly and annual comparison data from 4/4/2020 - 4/10/2020 @ 8:07 am PST 4/8/2020
Note: Averages have a *; Comparative Analysis and Pneumonia are highlighted in blue; COVID-19 is highlighted in Yellow


Weekly





Annual


Analysis

As of mid-week 6:
  • COVID is around 3.5 times more deadly than the High Flu peak weekly deaths benchmark at the mid-week point.
  • COVID is the 3rd most prevalent form of death in the US based on averaged weekly benchmarks at the mid-week point. 
  • COVID deaths have increased by around a factor of 2.2 since mid-week 5 (N = 2,572)
Corrections

After last weeks post the source data for mortality data made a correction in their totals due to a few reporting errors from states.  No large metric needs to be adjusted from last week; however, the total cases for the week ending 4/3/2020 should be revised to 5,826 which is a reduction of 277 deaths. 

These revised totals are reflected in the data moving forward. 



Saturday, April 4, 2020

COVID-19 Deaths as compared to other causes of Death (US Only) Week 5 Week update

Below is weekly and annual comparison data from 3/28/2020 - 4/3/2020 @ 8:52 AM PST 4/4/2020
Note: Averages have a *; Comparative Analysis and Pneumonia are highlighted in blue; COVID-19 is highlighted in Yellow


Weekly



Annual



Analysis, Benchmarks, and Factor Analysis

As of the end of Week 5:
  • COVID is around 3.8 times more deadly than the High Flu peak weekly deaths benchmark 
  • COVID is the 3rd most prevalent form of death in the US based on averaged benchmarks
  • COVID deaths have increased by around a factor of 5.6 since Week 4
  • COVID surpassed the annual mortality rate of HIV/AIDS in the US
Factor Analysis - below is a 'factor' analysis for COVID annual deaths versus benchmarks.  

Numbers which COVID does not exceed will be presented as X%.  This denotes that the COVID mortality rate equals X% of the higher number.

Numbers that COVID does exceed will be presented as N.  This denotes that the COVID morality rate is N times higher than the benchmark number.  A one (1) means that the numbers are equal, a two (2) that COVID is twice as deadly on an annual basis, etc.

For example, in today's analysis, COVID currently has a mortality rate 1.3 times higher than the estimated annual HIV/AIDS mortality rate.  To date, COVID has a mortality rate of  32.2% of the number of people who die annually during a Low-Moderate flu season.

All numbers are rounded to the nearest decimal.

Factor Analysis



Other info

A quick note on sourcing.  There are a few sites currently reporting COVID numbers.  Each lists a different number for infections and deaths.  This is due to a few factors primarily different time cut offs and different counting methods.  I have chosen https://www.worldometers.info/coronavirus because it has been the most consistent with updates and has the most granular data.  That said, you should treat each number as an estimate with an undetermined error range (probably in the neighborhood of +/- 50 - 100).

Finally a slight adjustment was made to the annual numbers for COVID this week adding 16 deaths that happened prior to the data set start.

The next update will be by 6 pm PST on 4/8/2020.