The Covid-19 graphs have been made using the dataset provided by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University at their GitHub page. The data used end on May 15, 2020 for the United States. In addition to the state and federal graphs, I want to include a project that I have been working on.
I have been working on processing the data from each county in the United States to show whether they are suffering from an outbreak of Covid-19 or not. Last Wednesday was the first day that I posted preliminary data from each county. Since then, I have honed the math and I believe that I can create a better picture of which counties should be monitored closely for Covid-19 outbreaks.
One of the main changes to the algorithm was to separate “hotspots” from outbreaks. I am using the term hotspot to indicate that there is at least a 7.5% rise in cases in the county resulting in at least 15 new Covid-19 infections. Hotspots, being relatively small increases, should be easier to control but would indicate that attention should be paid. On the other hand, hotspots could also be misleading since it would be easy to reach such a low threshold from testing while not necessarily indicating that Covid-19 is spreading as rapidly as indicated.
Hotspots are also unique in the sense that a county might start off as a hotspot, grow into an outbreak, get the outbreak under control, and pass through the hotspot stage again. Because of this fact, hotspots should be looked at closely instead of assuming that they are in the process of becoming outbreaks themselves.
Outbreaks are calculated virtually the same as I had calculated them in the past. They require at least a 15% rise in cases in the county resulting in at least 50 new Covid-19 cases. This metric seems to indicate that the outbreak has escaped the typical controls that are in place for the given county. While an outbreak could be indicated as a side effect of heavy testing in an area, it is much more likely that it would be indicated by unrestrained community spreading of the virus.
As always, here are a few things to keep in mind while you are looking over the results presented here:
- Each graph covers the dates from March 1, 2020 through May 15, 2020.
- Daily reported infections are recorded in red.
- Normalized infections are recorded in blue.
- The x-axis indicates the date.
- The y-axis indicates the number of Covid-19 infections (reported or normalized) on that date.
- Each y-axis is best fit. While this makes it easy to view the overall trend for the state, care must be taken when making comparisons to note the actual number of cases between different graphs.
- Negative numbers should never be present and are no longer shown on the graphs. Negative numbers would normally indicate a reporting error. One reason to ignore them is that they are small enough to not significantly effect the data presented.
With this new data in mind, here are the results for the week:
Continue reading Covid 19 update as of May 15, 2020
Marisa Burnes (dot) com 