Exploratory Data Analysis: Typical Occupations and Work from Home Habits in the CovIdentify Cohort

Ethan Ho, Andrew Brown, and Peter Cho

The CovIdentify study aims to use mobile and wearable devices to learn how we can detect early signs of COVID-19 and to predict whether individuals infected with COVID-19 will improve or worsen over time. Our study team is grateful to our study participants for donating their time and data to support this research. We have heard your feedback that it can feel like your data goes into a "black hole" where study participants do not know what happens next with their data. Because the scientific peer-review process can take a long time, (although it is necessary to produce rigorous science) as a study team, we feel strongly that we want to provide our participants with some of the information that we are learning as we analyze the data. Therefore, we are posting here the first of what we expect will be several blog posts demonstrating our initial findings from this study.

The figures in this blog post are based on the enrollment survey information that participants provided when signing up for the study. We hope to shed light on how the pandemic has changed our lives.

Work Status

The pie chart below shows the work from home status of people before and after the stay-at-home order was implemented in many locations around the US. (roughly March 16th, 2020). Approximately 75% of the participants in our study did not work from home before COVID-19. Of these, approximately 68% began working from home after stay-at-home orders were implemented. One-quarter of our study participants continued to work away from home even after the stay-at-home orders, indicating that roughly 25% of our study population may have essential occupations (see Figure 3 for a breakdown of occupation and work status).

Figure 1: Work-From-Home Status before and after March 16, 2020, when many areas implemented stay-at-home-orders.


The treemap chart below shows the breakdown of professions in the enrollment survey. We adapted occupational groups from the Bureau of Labor Statistics and divided professions into 8 categories (italicized throughout the post): Academic, Business & Banking, Engineering & Tech Design, Entertainment, Healthcare Profession, Service, Transportation & Labor, and Individuals Out of the Labor Force. Subcategories and details of each category can be seen by hovering over the graph.

The Healthcare Profession, Academics, and Retiree comprise a significant portion of the participants, whereas, there is little representation from participants working in the Transportation & Labor and Service sectors. One of our ongoing goals is to increase the representation of participants from these sectors.

Figure 2: Study participant occupations.

Work Status in Comparison to Profession

This stacked bar chart illustrates the breakdown of work status and profession. It allows us to see which professions continued to work away from home after the COVID-19 outbreak, until the most recent analysis carried out on 5/26/2020. For each bar, the colors signify the percentage of a particular profession category in one of the four work-from-home categories: worked from home only before, worked from home before and after, worked away before and after, and worked from home only after.

Professions that neither worked from home before nor after the outbreak (in turquoise) may indicate those who might have essential occupations, where their occupations require them to continue working away from home amidst the pandemic. Healthcare, Transportation & Labor, and Service professions have a larger proportion of individuals continuing to work away from home.

Figure 3: Work-From-Home Status by Occupation.