COVID-19 Mortality & Socioeconomics in Michigan Counties

updated Aug 20 2022

Listen to sufjan stevens My source code is on github.
Note: This is a work in progress, and more analysis is needed to compute the statistical significance of the findings.

Project Description

My script automates the following:

  • Collects and aggregates the most recent cumulative count of COVID-19 deaths for Michigan by county
  • Calculates COVID-19 deaths per thousand
  • Groups the 83 counties of Michigan into 5 risk segments (quintiles), Q1 being the lowest rate of COVID-19 deaths
  • Connects to the census.gov database and retrieves the latest population estimates
  • Compiles the unemployment rate, education rate, household public assistance utilization rate, median income, median age, and county racial compositions
  • Joins the two datasets together via county
  • Reveals the summary statistics of the numeric variables
  • Vizualizes the data into tables and scatter plots

As of August 20th 2022, this data reflects COVID-19 deaths as of August 19th 2022

Sources: American Community Survey (ACS) 2016 - 2020 Estimates, and the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University

Background and Bias

Introduction to COVID-19 Mortality & Socioeconomics in Michigan Counties

At the age of 16, my grandmother “Nana” left her parents and nine siblings behind in pursuit of a better life in Grand Rapids, Michigan. Her parents were sharecroppers, and if you know anything about 1950’s Mississippi – life wasn’t exactly ideal for a young black woman.

Upon moving north, Nana had the good fortune of meeting Betty, who was secretly dating a well-to-do judge. Betty’s boyfriend bought her a house in a white middle-class neighborhood… This was a HUGE scandal. You see, blacks were not “allowed” to own property in that part of town. Betty’s boyfriend found a loophole. According to my grandmother, the neighbors were completely beside themselves, and promptly vacated the area.

My grandmother bought the home next to Betty for a fraction of the cost by the distraught homeowners. By the time my father was born, the community shifted into a black middle-class West Michigan neighborhood. In the 1980’s, a large public housing project was proposed and constructed next to my grandmother’s home, and the community evolved again. Nana still lives here today.

Acknowledging Bias

I share this story to disclose my potential bias that may shape my analysis of the 2016-2020 American Community Survey (ACS) and the COVID-19 data provided by JHU CSSE. Like many, I understand how a zip code profoundly shapes quality of life. As I built the scatter plots and summarized the variables, I did my best to let this notion go and let the data speak for itself.

Why this topic?

The global collaboration to document and report the status of the pandemic over the last few years is truly remarkable. However, the longer the quarantine went on, the more I focused inward.

During the pandemic I:

  • Returned from a year abroad
  • Ended my 8 year military career
  • Got married
  • Had my first child
  • Bought a house
  • Enrolled in grad school
  • Changed careers

These factors do not excuse my ignorance, but perhaps will shed light regarding my topic choice.

While I had a peripheral understanding of COVID-19’s devastating impact on the world at large, I was blind to many of the state-wide implications. This project is an attempt to rectify my knowledge gap, and inform others in a similar position.

Throughout this process I learned so much about Michigan! In terms of demographics, Michigan counties vary widely. Admittedly, I was pretty ignorant of my home state.

I hope my work provides you with a new perspective into Michigan counties.

Guide to the Tables

  • Summary Statistics: Use this as a point of reference when considering “How does this county compare to the rest?”
    • For example:
      • Does a county skew young or old?
      • What’s the average unemployment rate in Michigan?
      • What is the highest county public assistance rate?
  • All the Michigan County Data:
    • Here you will find a list of all MI counties and their key socioeconomic figures side by side
    • You can search, filter, and sort my entire county-COVID dataset
  • COVID-19 Mortality Quintile (Q1 through Q5):
    • This final table allows you to compare the socioeconomic differences between a Q1 county, a county with the lowest COVID-19 mortality rate, and the highest rate (Q5)

{jump to the interactive table} screenshot

The Findings:

With an average of 2.4 deaths per thousand, counties in the first quintile (Q1), look radically different than the other end of the spectrum:

Traits of the average Q1 MI county:

  • medium sized populations
  • younger residents
  • highly educated residents
  • the highest median income of all the quintiles
  • racially diverse
  • the least amount of unemployment
  • the least amount of households on public assistance

Example Q1 counties include:

  • Kent
  • Alger
  • Grand Traverse
  • Kalamazoo
  • Marquette

The Relationships Visualized

The interactive graphs show the relationship between one socioeconomic variable and the COVID-19 mortality rate. Hover over a plot to learn more about the county. The larger the bubble, the more deaths that county experienced. The higher the quintile, the more death the county experienced per 1000.

COVID-19 Mortality by Median Age

full screen view

r = 0.49 | moderate positive relationship: as median age increases, the COVID-19 mortality rate increases

The figure shows that Otononagon, a Q5 county, has the highest median age of 59.1.

As of August 19th 2022, 40 people died of COVID-19 (7.0 per thousand) in Otononagon county

COVID-19 Mortality by Median Income

full screen view

r = -0.50 | moderate negative relationship: as median incomes increase the COVID-19 mortality rate lessens

All of the Q5 quintiles fall under a median income of $60k, with most counties’ income falling in the range of $40-50k. Livingston, a Q1 county with a median income of $84k just surpassed Oakland, a Q2 county, as the wealthiest area in Michigan.

As of August 19th 2022, 479 people died of COVID-19 (2.5 per thousand) in Livingston county.

COVID-19 Mortality by Higher Education Attainment

full screen view

r = -0.50 | moderate negative relationship: as percentage of education attainment increases the COVID-19 mortality rate lessens

Home to the University of Michigan, Washtenaw is Michigan’s most educated county, with 56.7% of adults 25+ holding college degrees. This Q1 county stands out considering most Michgian counties rate of higher education attainment is under 30%.

As of August 19th 2022, 554 people died of COVID-19 (1.5 per thousand) in Washtenaw county.

COVID-19 Mortality by Public Housing Assistance Rate

full screen view

r = 0.37 | weak positive relationship: as rates of public assistance increase, the COVID-19 mortality rate increases

Wayne, the most populous and a Q4 county, stands out as the area with the highest percentage of households on public assistance.
As of August 19th 2022, 8,267 people died of COVID-19 (4.7 per thousand) in Wayne county.