COVID-19 Data and Models for the United States


Updating Data . . .

Disclaimer: I am not an expert in epidemiology. I am a mathematician that desires more information about COVID-19 than has been available.

The short version of what the data below shows: Go straight to the Discussion.

Quick Links To Resources On This Page

Current and Past Data

Vaccinations To the U.S. data we have added World data and country data where available.
U.S. State and the County of Your Choice The states and counties now have additional hospital data from the department of Health and Human Services (HHS).
Rank The U.S. States (Worst to Best) New: Rank vaccinations.

Sources and Notes About This Site

Data for this site is from https://github.com/CSSEGISandData/COVID-19/tree/master/csse_covid_19_data/csse_covid_19_time_series which is the raw data that Johns Hopkins makes available.
Vaccination data is from https://github.com/govex/COVID-19/tree/master/data_tables/vaccine_data which is the raw data that Johns Hopkins makes available.
The covidtracking.com website has decided that the CDC and HHS are now doing a credible job so covidtracking.com will stop collecting data on March 7. Now all Testing and Hospitalization data is from the Department of Health and Human Services at https://beta.healthdata.gov/dataset/COVID-19-Community-Profile-Report-County-Level/di4u-7yu6.
Population data is from the United Nations Department of Economic and Social Affairs using their medium level projections for 2021. https://population.un.org/wpp/Download/Standard/CSV/.
Other useful sites are

Johns Hopkins website: https://www.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6

http://91-divoc.com/pages/covid-visualization/

If you are curious to play with an epidemiologist's modeling tool that uses the traditional R nought, etc, see https://covid19-scenarios.org/. You will see that it's predictions are extremely sensitive to very slight changes in the assumptions which is why it is so difficult to make predictions and also why not using a mask to protect others, not social distancing to protect others, and not taking the pandemic seriously is foolish.
Note: Death rates on this site are computed using current deaths divided by cases 14 days ago. Using current cases erroneously assumes that everyone that currently has covid lives.
Note: The New Tests Per 100K Goals are based on testing enough to move the positivity rate towards its goal. For example, in the U.S. if the positivity rate is at 10% and new tests per 100K is at 500 then doubling the number of new tests to 1000 per 100K would give a positivity rate of 5% if the additional tests all came back negative. Since some of the additional tests would be positive, it means that even a testing level of 1000 tests per day per 100K is still not sufficient. As in this example we use this multiplier of positivity/5 times the current number of new tests per 100K to get a goal. The same idea for new tests is used for a 3% positivity rate goal.

Vaccinations - U.S.


Vaccinations - The World

Sort the countries below and their vaccination numbers and rates either alphabetically, by number of vaccinations, by percent of vaccinations for two per person, or by population. The data is incomplete, but these are the countries for which Johns Hopkins has the vaccination data.


If We Had Made An Effort

Why highlight the failure of the U.S. below? Because if we learn from our failure we can still save hundreds of thousands of lives while the vaccines are being distributed.

Genome Sequencing

Genome sequencing allows us to trace the spread of COVID-19 and to keep abreast of the mutations as they occur.
Update: January 23, 2021. Denmark is now sequencing every new case. The U.S. is still flying blind and has no idea of what is happening. Experts say at least 5% of cases need to be sequenced. As of late January, the U.S. ranks 38th out of 130 countries reporting whole-genome sequencing data. See Why America is 'flying blind' to the coronavirus mutations racing across the globe. Researchers warn the U.S. desperately needs to sequence more genomes so it can stay ahead of new variants
Update: February 12, 2021. This expert says we should sequence 25% to 30% of positive COVID-19 tests. See How to Track Viral Variants. The emergence of potentially more dangerous SARS-CoV-2 variants highlights the urgency of viral surveillance efforts.

Current and Past U.S. Data


New Cases and New Deaths in the U.S. in the Last Eight Days

These are daily values which fluctuate on weekends due to reporting delays as opposed to 7-day averages which are below.

7-Day Average of New Cases and New Deaths in the U.S. in the Last Eight Days


Cases By Month in the U.S.



Deaths By Month in the U.S.

Percent to increase (or decrease) reported COVID-19 deaths: %
Deaths are likely underreported. See True Pandemic Death Toll in the U.S. 34% higher than reported (Dec. 5). Also see the CDC report which suggests Deaths are 42.6% higher than reported: Excess Deaths Associated with COVID-19. Adjust a percentage increase (or decrease) for COVID-19 deaths as desired. The above sources indicate that deaths are likely 34% higher or 42% higher than reported so the default is the average of 38%.

Cases and Deaths for the Country of Your Choice

The WHO recommends keeping positivity below 5%. If too high, the sickest patients are found, but milder or asymptomatic cases are missed. This means flare-ups may be missed and get out of control. The Harvard Global Health Institute recommends below 3% for suppression as opposed to just mitigation. See here.
Country: Days to average new cases and new deaths:

Cases and Deaths for the U.S. State and the County of Your Choice

State: County: Days to average new cases and new deaths:

Rank The U.S. States (Worst to Best)

The coloring is a range of red (worst) to green (best) of all states and D.C. All computations involving new cases or new deaths use the last 7-day average so that irregularities concerning weekend reporting are evened out.
New Cases Change and New Deaths Change have a range of red colors for positive percent changes (bad) and greens for negative percent changes (good).
The Goals listed below on the right in red are for working towards mitigation and those listed in green are for working towards suppression.
The goal of greater than 150 new tests per 100K has been removed since that is no longer a reasonable goal when new cases are high and positivity rates are well over 5%. For example, in the U.S. if the positivity rate is at 10% and new tests per 100K is at 500 then doubling the number of tests to 1000 per 100K would give a positivity rate of 5% if the additional tests all came back negative. Since some of the additional tests would be positive, it means that even a testing level of 1000 tests per day per 100K is still not sufficient.
States:
Choose How to Rank:















100% means that the current new cases (deaths) are matching or exceeding the previous peak.









The WHO recommends keeping positivity below 5%. If too high, the sickest patients are found, but milder or asymptomatic cases are missed. This means flare-ups may be missed and get out of control. The Harvard Global Health Institute recommends below 3% for suppression. See here.




Vaccinations


No longer have data for # people with one shot or two (Some states have changed how they provide their data in inconsistent ways).


Up to one point for each goal (using the red mitigation goals):
1)  New cases per 100K. Goal: < 4 or < 1.
2)  New cases 2-Week Change. Goal: < 0.
3)  New Tests Positivity Rate. Goal: < 5% or < 3%.

This gives some measure of how the states are currently doing while past results and total numbers of cases has no effect on this score. The best possible score is 3.


The U.S. Compared to the European Union


Groups of States


Expert Quotes, Discussion, and Links

Expert Quotes

Three random quotes for the day or all quotes.

Discussion

Even with the vaccines that we have, it will take until September to reach herd immunity and that assumes that the vaccines can be adjusted to handle coronavirus variants that arise. Many, many thousands of people will still die needlessly in the next year and millions of people will suffer long term COVID-19 effects. Even if you can rationalize the immorality of these needless deaths and think we should just wait for the vaccines to take effect, we do not yet know how well it will work. A recent study (Immunity to Covid-19 could be lost in months, UK study suggests) shows that we may just get COVID-19 again in a few months. A vaccine might require regular boosters. We also do not yet know if vaccinated people can still spread COVID-19. The South African variant is both transmitted more easily and is more deadly. This and other mutations of the coronavirus will require that we keep updating the vaccines.
We must stop thinking only of ourselves, start thinking of the welfare of others, be responsible, and rise to meet our obligations to society. To do this we must develop toughness and determination. We must do what other countries have done. We must shut down everything that is nonessential, everyone must wear a mask, and everyone must aggressively social distance (What we have done instead has been far more damaging to the economy and to people's lives). For people to be able to do this and survive, we must provide living expenses, as many other countries have done, for everyone during this time so that people are not forced to break quarantine. The result will be that instead of failing to save either lives or the economy we will in a matter of months save both.
During these months we must learn about the effectiveness of the vaccines, build the infrastructure to test, contact trace, and genome trace. A genome sequencing program will give us information about the mutations of the coronavirus in our country early on before they spread, instead of learning about them from other countries after they have spread to them. Such a delay means that we cannot react to mutations with vaccine adjustments early in the process and more people will die. We will have to remain vigilant and the battle will continue. Our lives will change and we will not go back to how things used to be, but we will have lives and we will be able to rebuild. This is not that hard to do and it is our moral obligation. We have had no difficulty in providing a trillion dollars for corporate wellfare and no one questioned how it would be paid for. Why is a similar effort for saving lives too costly? To paraphrase Martin Luther King Jr., our country loves socialism for the rich and the harsh injustice of capitalism for everyone else.
The longer we keep delaying real action, the greater the number of needless deaths.

Listen to the epidemiologists: Wear a mask and aggressively social distance. Don't rationalize your behavior (going to events, vacations, etc) by saying that your mental health is more important than your physical health. Yes, mental health is important, but you dying, your family members dying, your friends dying, or you causing the deaths of random strangers which you may not know you have caused, is worse than mental health problems. 40,000 children in the U.S. have lost one or more parents to COVID-19.

Percent of wages subsidized by governments due to COVID (December 2020):
  • Japan: 100% for small businesses; 80% for large firms
  • Netherlands: Up to 90%
  • Norway: Up to 90%
  • Germany: Up to 87%
  • France: Up to 84%
  • Italy: 80%
  • United Kingdom: Up to 80%
  • Canada: Up to 75%
  • United States: 0%
5 Steps for Viral Surveillance. See How to Track Viral Variants
  • Submit a portion of positive tests-Andy Pekosz (Professor in Molecular Microbiology and Immunology) says 25%-30% is ideal-for genomic analysis.
  • Compare these sequences to existing ones stored in databases like GISAID and CDC's SPHERES to spot changes.
  • Track mutations-especially those related to the spike protein-that begin appearing frequently.
  • Investigate whether those mutations affect the efficacy of tests, therapeutics, or vaccines.
  • Design tests that rapidly detect variants of particular importance.

Additional Links and Information


Personal Stories

To gain an understanding beyond the data it is important to see and feel what people are going through. Below is the beginning of an ongoing collection of personal stories related to COVID-19.

Prediction Resources

Damage of Heart, Lung, Kidney, Brain, Triggering Diabetes, Hearing loss, and Other Serious Consequences of COVID-19

Very little is known about the long term effects of COVID-19. That does not mean that we should ignore what we do know. Many lives and the quality of those lives depend on the decisions that we make now. The fate of our economy also hangs in the balance.
We won't have a complete picture of the long term effects of COVID-19 for many years, but some studies are starting to give us a glimpse of what we can expect. The chart below shows ranges for the numbers of people with damage to their heart, lungs, kidneys, brain, triggering diabetes, hearing loss, and general problems based on studies available to date. The sources for the ranges of these damages, and general problems are cited below and will be updated as more information becomes available.

Select the Number of Positive COVID-19 Cases in the U.S.


Predict Cases and Deaths for Six Months

Technical Details: The last two weeks of rates of changes of cases or deaths is approximated as an exponential function (using least squares fitting) which is used to predict the future cases or deaths for six months. Additionally, since some countries and states are experiencing increasing exponential growth, after 10 days if cases reach 2% of the population or deaths reach a tenth of that, then the new cases or deaths per day are held constant. The graphs clearly show when this happens as a horizontal line off to right. A seven-day rolling average of the new cases and new deaths is used in order to even out irregular reporting over weekends.
Warning: Recall George Box - "All models are wrong, but some are useful". Do not think about these predictions in terms of the resulting numbers. Do think of these predictions as trends which show where we may be headed. Also note that we cannot rely on some of the data. Some countries and states are both exaggerating testing results, by counting multiple positive tests for the same person and counting antibody tests as positives, and underreporting deaths as has been widely reported in the U.S. If Predicted CFR is much less than Current CFR there may be underreported deaths. Not applicable to the U.S., but other reasons that the CFR may be lower is if widespread random testing is being done so that many positive asymtomatic cases are found or if a country or state is successfully contact tracing so that a lower percentage of those that catch COVID-19 are the older people at highest risk. This saves lives and lowers the CFR. If the numbers of cases and deaths are low then the predictions are not likely useful.

State:
Country:

Sorted Graphs for All States of either New Cases or New Deaths

The previous system of graphing the light blue graph to help spot when states or countries are cheating by underreporting deaths and overreporting cases is now obsolete because over time it has become clear that the first two points listed below have led to a real decline in the the case fatality rate (CFR). A new method is described below.
When graphing New Deaths with no predictions a light blue graph is also plotted. It is the graph of new cases multiplied by the death rate. There are also two additional adjustments - It is moved to the right or delayed 14 days and the last 14 days are removed. These two changes are made because the typical time from contracting the coronavirus until death is about two weeks. This is useful because if the case fatality rate (CFR) is constant, then this graph would closely match the new deaths graph which is also shown. This gives a way to visually see how the CFR has changed. The typical situation is to start out in the 5% - 10% range and currently be in the 1% - 3% range. On the graphs this means that the light blue graph will be a little below the dark blue graph on the left side and above on the right side.
There are reasons that CFR has changed over time:
1) Treatment has improved.
2) We are testing more people with either minor or no symptoms.
3) States and countries are distorting the data. Nine states mix COVID-19 testing with antibody tests. Some states and countries have adjusted and changed how they define COVID-19 deaths. Some states have been caught directly altering the data to make it look more favorable.
It is difficult to know how legitimate each state's or country's data is because of the mixed effects of the above three factors along with many others. But generally speaking, if the light blue graph is not a little below the dark blue graph on the left side and above on the right side then we should ask why.
Below each graph the CFR percent is shown for five consecutive 4-week periods ending yesterday.
Graph either with (slow) or without predictions:

Sort as follows:


Sort worst to best as follows:






Current Overall Performance adds up to one point for each goal:
1)  New cases per 100K. Goal: < 4
2)  New cases 2-Week Change. Goal: < 0.
3)  New Tests Positivity Rate. Goal: < 5%

This gives some measure of how the states are currently doing while past results and total numbers of cases has no effect on this score. The best possible score is 3 (at the bottom of the list).

Sorted Graphs for All Countries of either New Cases or New Deaths

The same note that is above for the states concerning graphing New Deaths with No predictions and the light blue graph that is also plotted also applies for the countries below.
Warning: There are 214 countries so these graphs take some time to load (7 seconds for me. With Predictions, 10 seconds.)
Show Countries With Populations Between and

Graph either with or without predictions:

Sort as follows:


Sort worst to best as follows:







Testing data is not available for countries so still using the following version of Overall Performance for countries which counts, with equal weighting:
1)  Cases per 100K,
2)  New cases per 100K, and
3)  New cases per peak new cases. So countries that have kept new cases low are not penalized too much, this factor is adjusted downward if new cases per 100K is less than 1, in which case we multiply by new cases per 100K.

The same is done for deaths except the third factor is adjusted downward if the new deaths per 100K is less than 0.05, in which case we multiply by new deaths per 100K divided by 0.05.

This gives some measure of how the countries have done and are currently doing. The highest or worst possible score is 100 (at the top of the list).

Additional Resources

COVID-19 Deaths in the U.S. Compared to Other Causes of Death

There are many reports in the media comparing deaths by COVID-19 to deaths by other causes. Some of these comparisons use different time frames which skews the results. The following chart uses data from the CDC for 2018 deaths (2019 is not available yet). The data is used to compute average deaths per day so that we can compare to the average COVID-19 deaths per day. For the COVID-19 average we use deaths since April 1, 2020. Before April there were 5,605 COVID-19 deaths out of hundreds of thousands now so it had not really gotten going yet.
Percent to increase (or decrease) reported COVID-19 deaths: %
Deaths are likely underreported. See True Pandemic Death Toll in the U.S. 34% higher than reported (Dec. 5). Also see the CDC report which suggests Deaths are 42.6% higher than reported: Excess Deaths Associated with COVID-19. Adjust a percentage increase (or decrease) for COVID-19 deaths as desired. The above sources indicate that deaths are likely 34% higher or 42% higher than reported so the default is the average of 38%.
Remark: COVID-19 is around 10 times worse than the flu. This does not take into account that recovery from the flu usually means complete recovery and there is increasing evidence that this is not the case with COVID-19. We don't yet know how long heart, lung, kidney, brain, hearing, and other problems will persist after COVID-19 and shorten lives.
Remark: Unlike cancer and heart disease, over 99% of COVID-19 deaths were preventable just as many hundreds of thousands of future deaths can still be prevented.

COVID-19 Deaths in the U.S. Compared to Wars and the 1918 Pandemic

Percent to increase (or decrease) reported COVID-19 deaths: %

Example of Runaway Exponential Growth