Biometrics Northwest LLC

Performing Data Analysis and Modeling

Home

Services

About Us

Projects

Contact Us

Disclaimer

Project: Dynamic Covid-19 visualizations (a.k.a movies)

I had not seen any dynamic visualizations (a.k.a. movies) of the spread of Covid-19 in the U.S., or the world for tha matter, so I made state and county level visualizations for the U.S. using the data from two sources.

  1. Risk Levels - Global Epidemics: Key metrics for covid suppression maintained by the Brown University School of Public Health and Microsoft's AI for Health program.
  2. The COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University: https://github.com/CSSEGISandData/COVID-19. Specifically, the data used were the time series data for confirmed Covid-19 cases.

The metric used for determining the risk class was number of new daily Covid-19 cases per 100,000 people and there are four risk classes: good, spreading, fast spread, and bad (my name choices). A rationale for this metric can be found here.
Brown University School of Public Health - Microsoft AI for health
 Covid-19 spread by U.S. State Starting date: January 21, 2020
 Covid-19 spread by U.S. County Starting date: January 1, 2021

The COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE)
 Covid-19 spread by U.S. State Starting date: January 21, 2020
 Covid-19 spread by U.S. County Starting date: January 21, 2020

 The Johns Hopkins data were added since they had matching starting dates for the state and county levels: January 21, 2020, and hence provide a longer timeframe for viewing the spread of Covid-19.

For these visualizations states or counties not reporting data are shown in a gray color. In addition, states or counties having a metric value of exactly zero on a particular date are also shown as gray. This was done for three reasons.

  1. The Brown University/Microsof AI for health metric data available for download were rounded to two decimal places, so low numbers of new Covid-19 cases per 100,000 people could have been rounded down to zero. These would still have been in the good category, but it is not possible to distinguish a rounded to zero value from the following two situations.
  2. Zero values could have been reported by a state or county because they are no longer tracking new cases and reporting actual case counts.
  3. A zero value could have been used as a default value for the numeric data during processing, so a state or county that was not reporting could have been assigned a zero value for the metric (this may be the case for most counties in Nebraska).

Back to top

For information send email to: info@biometricsnw.com

Last Update: September 22, 2021 11:00 AM

Copyright 2005-2021 Biometrics Northwest LLC