I have been tracking the COVID-19 pandemic since the Detroit metro area was hit hard by the initial spring 2020 outbreak. I live in southeast Michigan, and the pandemic has yet to end, so I compile the indicators I watch most closely into a single dashboard, which I host at https://covid19.timothystehulak.com.

I want to know: Are more people getting sick? Are outbreaks being contained? Are we doing enough testing? Is more testing “causing” more cases, or are new cases outpacing testing? In this post, I will describe the indicators I am showing on the dashboard. In a future post, I will describe the technology used to create it.

Geography

The dashboard is designed around geography. The concept is straightforward: let the user select a location from a dropdown menu, and then show all relevant charts for that area. It takes data from the State of Michigan coronavirus website as an input and focuses only on six counties in southeast Michigan.

Counties are the smallest subdivisions I consider, and I follow the state convention to report the City of Detroit separate from the rest of Wayne County. Some data are available at the healthcare coalition region level (which is a collection of counties), and I show that regional data when county data are not available. In general, healthcare data are available at the regional level, while testing and vaccination data are available for each county.

Healthcare coalition Region 2 South includes Monroe, Washtenaw, and Wayne counties; Region 2 North includes Macomb, Oakland, and St. Clair counties. “Region 2” is an aggregate I made as the sum of both regions. This aggregate can be thought of as the metropolitan area. Although it does not fit the Office of Management and Budget definitions for the Detroit-Warren-Dearborn MSA or Detroit-Warren-Ann Arbor CSA, it falls somewhere in the middle of the two. The “Region 2” aggregate misses Lapeer, Lenawee, Livingston, and Genesee counties from the CSA. Those counties are in either Region 1 or 3, but those regions are large and contain areas well beyond the CSA, including Bay City, Flint, Jackson, Lansing, Midland, and Saginaw.

Infection Charts

Most of the charts I show about COVID-19 infections are standard charts for understanding the pandemic. The number of cases is reported by date of first symptom, so newly identified cases may be attributed to days prior to the day test results are reported. This is different from the number of positive tests, which is used as an indicator of when positive cases are recorded; this is a near real-time indicator of disease prevalence. Another important distinction between cases and tests is that an infected person is only counted as a case once but may be tested more than once.

Test positivity is the percentage of tests which are positive on a given day. Test positivity is not necessarily a good indicator of disease spread because it is based on a non-random sample of individuals that voluntarily get tested.

The two-week moving average of tests shows the total number of tests and the breakdown between how many are positive or negative. The moving averages smooth volatile daily data.

The most novel chart I made shows the growth rate of positive tests vs total tests. In the summer and autumn 2020, I heard claims made that testing creates more cases, or that cases are up because testing is up. The last chart shows there is clearly more testing now than in spring 2020 when tests were scarce. The growth rate chart shows the how much testing has grown in the most recent two weeks vs the two weeks prior. If the growth rate of positive tests exceeds the growth rate of total testing, then an increase in testing is not completely explaining the increases in positive tests. In other words, this could indicate a real increase in disease prevalence or possibly an outbreak.

Regional outbreaks and hospitalization

Data on outbreaks and hospitalizations are shown for healthcare coalition regions, but not counties. This data is available from the state website. It is reasonable to show because outbreaks may spread throughout a region, and infected individuals may use resources in the region that are not necessarily in county they reside.

Two charts are about outbreaks, which are times when individuals from different households are in the same setting and become infected. New outbreaks indicate new investigations by local health departments, while the total includes new and ongoing outbreaks.

Hospitalization trends are important because hospitalizations are for individuals with more serious symptoms and because hospitals can become overburdened if there are too many ill patients. The most serious cases are shown by the number in the ICU and on a ventilator. Finally, I show the percentage of visits of individuals with COVID-19-like symptoms who visit select medical centers for care.

The state reports excellent vaccine data. Charts indicate how many of each type are sent to medical providers in the area, and how many of each type have been given to residents. It is possible for an area to have inoculated more residents than it received because people cross county lines to get the vaccine.

Deaths

This pandemic is a tragedy, and every life lost to COVID-19 mattered. The final chart shows both the number of daily deaths and the cumulative death toll. Sadly, the total moves higher as more individuals die succumb to the disease.

View the dashboard

The COVID-19 dashboard I made tries to clearly lay out some of the most important indicators of the pandemic for southeast Michigan. Further descriptions are available on the dashboard itself. If you would like to view the dashboard, it is available at https://covid19.timothystehulak.com. The next post will describe the technology behind the dashboard.