US COVID-19 cases

Linear Extrapolation

The author's simple linear extrapolations of the COVID19 confirmed cases and deaths for the US between August 1st 2020 and September 10, 2020 from the Johns Hopkins University (JHU) COVID-19 database yield almost 300,000 deaths and 11.3 million confirmed cases by the end of 2020:

Deaths

Confirmed cases

Spreadsheet

covid-us-predict.xls

University of Washington pandemic forecast

See http://www.healthdata.org/sites/default/files/files/Projects/COVID/briefing_US_091120.pdf

This looks at the current state and the trends in key drivers of transmission (mobility, mask use, testing, and seasonality). They expect the daily death rate in the US, because of seasonality and declining public vigilance, to reach nearly 3,000 a day in December. Cumulative deaths expected by January 1 are 415,090; this is 222,522 deaths from now until the end of the year; or over 120,000 deaths more than the simple linear extrapolation above.  The following comments from the forecast are very cogent.

• "The large increase in daily deaths expected in late November and December is driven by continued increases in mobility, declines in mask use, and – most importantly – seasonality. We estimate the likely impact of seasonality by examining the trends in the Northern and Southern Hemisphere. For example, Southern Hemisphere countries such as Argentina, Chile, southern Brazil, and South Africa had much larger epidemics than expected based on mobility, testing, and mask use. The statistical association between COVID-19 transmission rates and pneumonia seasonality patterns is strong and is the basis for our estimate of the magnitude of the seasonal increase that is expected. • If a herd immunity strategy is pursued, meaning no further government intervention is taken from now to January 1, then the death toll could increase to 611,784 by January 1. Compared to the reference scenario, this would be 196,694 more deaths from now to the end of the year. 
• Increasing mask use remains an extraordinary opportunity for the US. Increasing mask use to the levels seen in Singapore would decrease the cumulative US death toll to 298,589, or 116,501 lives saved compared to the reference scenario."

The top 9 states are shown below ranked by their Positivity Rate (from https://coronavirus.jhu.edu/testing/testing-positivity 9/8/2020) together with some demographics and statistics. The smoothing is a 7 day moving average.

• The population is from https://en.wikipedia.org/wiki/List_of_states_and_territories_of_the_United_States_by_population,
• the political leaning is from www.270towin.com updated: Sep. 9, 2020
• the Income per capita is from  https://en.wikipedia.org/wiki/List_of_U.S._states_and_territories_by_income.

The following charts are obtained from the analysis of the Covid-19 statistics are from the Johns Hopkins University (JHU). There are raw data for confirmed cases and deaths by date for each county in each US state. For the US state analysis, the data from the counties are aggregated into the values for each state.

Cumulative confirmed cases Click on chart for more details	Cumulative deaths Click on chart for more details

Covid-19 % deaths/confirmed cases Click on chart for more details	Covid-19 confirmed cases per million population Click on chart for more details	Covid-19 deaths per million population Click on chart for more details

Covid-19 smoothed confirmed cases per day per million population Click on chart for more details	Covid-19 smoothed deaths per day per million population Click on chart for more details

Details about the timelines, government resposes to, and impact of  Covid-19 on the US states can be found via https://en.wikipedia.org/wiki/Category:COVID-19_pandemic_in_the_United_States_by_state. It is not hard to believe that some of the surges are related to events, behaviors, interventions, and responses. For example:

• In South Dakota, the surge of confirmed cases in July following the July 3rd Mount Rushmore Independence Day celebration. This in turn was followed by a spike in confirmed cases following the Sturgis motorcycle rally  August 7-16.

Observations:

• Florida reported its first cases 3/3/20 and its first death 3/8/20, the latter date being the same day that Kansas and Missouri announced their first confirmed cases. None of the other 6 states had any confirmed cases by this date.
• As would be expected since Florida has over three times the population of any of the other states, its confirmed cases and. deaths lead the way by a wide margin.
• In terms of deaths/confirmed cases, Missouri led until the end of July when Mississippi took the lead.
• When one normalizes cases by the population of each state, then the confirmed cases and deaths are more evenly spread, and none of the states has flattened its curve of confirmed cases.
• Idaho appears to have managed to flatten its curve of deaths/million population after 5/11/20 until 7/15/20 when it took off again.
• Florida, Mississippi, and Alabama appear to have a peak in daily confirmed cases two to three weeks after the July 4th holiday. There is a further rise in daily confirmed cases for North and South Dakota, Iowa, and Alabama towards the end of August.

Method

We are interested in whether some of these events may have served as 'super-spreader' events. 

The raw data we analyze is the Johns Hopkins University daily confirmed cases data for the US available at:  https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_US.csv

The question is whether there was a significant rise in daily confirmed Covid-19 cases after the rally, in particular in the 2-14 days following the rally (according to the CDC, Sep 18, 2020: "Symptoms may appear 2-14 days after you were exposed to COVID-19").

We have not analyzed the more recent rallies following President Trump's experimental treatment at the Walter Reed hospital for the virus, since there is insufficient delay compared to the 14 day incubation period, to clearly identify a rally's impact.

The social distancing at the rallies was minimal, and not everyone wore masks, see for example, the photos below taken at the Oshkosh rally.

Results

The links "Details in Excel file" in the table below (derived from https://en.wikipedia.org/wiki/List_of_post-election_Donald_Trump_rallies), provide access to Excel worksheets and charts of daily confirmed cases for each county plus identification of the event's starting date followed by a 14 day incubation period.

On clicking the "Details in Excel file" link it will prove access to the Excel workbook. On opening the workbook, open the Chart1 tab. You can move the mouse over a chart's line to get the date, value and county for the line.

The Excel files come from a library of Excel workbooks illustrating the impact of COVID-19 on many different areas of the world.

Possibly Impacted counties

We looked at the adjacent counties in the same state or neighboring state (using https://gisgeography.com/us-county-map/) where we see the following counties.

Cumberland NC 9/19/20	Montgomery OH 9/21/20	Fulton & Lucas OH 9/21/2020	Allegheny PA 9/22/2020	Duval FL 9/24/20	Newport News VA 9/25/2020

Also see

• https://time.com/5865890/oklahoma-covid-19-trump-tulsa-rally/
• https://www.axios.com/trump-rally-fined-violating-covid-guidelines-1566236b-0177-4407-b139-a7f69c11db66.html
• https://thenevadaindependent.com/article/nevada-health-officials-predict-uptick-in-covid-19-cases-following-labor-day-weekend-crowds-trump-rallies
• https://www.usatoday.com/story/news/investigations/2020/10/22/trumps-campaign-made-stops-nationwide-then-coronavirus-cases-surged/3679534001/