WMU librarian preserves COVID-19 data for future researchers

In the first few weeks of the pandemic, Western Michigan University librarian Daria Orlowska looked at the growing number of COVID-19 webpages and data dashboards and asked an important question: Who's saving the data when it is updated daily?

In the early months of the pandemic, state and local governments had to figure out what information the public needed and the best way to communicate it. Archiving webpages and dashboards may not have been a high priority when people needed the most up-to-date information to stay safe and make decisions for their families.

"I was particularly frustrated by the lack of aggregate [data], especially since each daily update replaced the previous day," Orlowska says. "And this made me wonder, how can we make this information easier to parse for future researchers?"

This observation led to an ongoing data archiving project that started in April 2020. Orlowska began to capture and compile the data in a downloadable file format that allows users to view and analyze it more efficiently and that preserves this historically relevant information.

"I had already been tracking data released by the Michigan Department of Health and Human Services when the pandemic hit Michigan at the beginning of March 2020 and was struck by the changes in data reporting," Orlowska says. "There was a lot of speculation about the state of hospital preparedness, so when the Statewide Available PPE and Bed Tracking webpage was first made public in early April 2020, I knew that aggregating released data and preserving the original webpages would be important."

Information on statewide hospital occupancy and availability of personal protective equipment (PPE) is tracked on the Michigan.gov COVID-19 website. But it is not publicly archived or available in a format that can be easily analyzed.

Her work also documents the changes made to variables tracked by the state, which will help researchers understand how reporting changed over time.

"During the first month of release, the variables changed a lot. Some hospital units and PPE [data] were captured one day, disappeared the next, only to reappear shuffled around with a name change," says Orlowska. "I decided on a set dataset structure and collapsed variations in name into the same column when I was confident that they were the same, documenting the name changes and dates they were in use."

Our Michigan COVID-19 Data collections are available to the public online through ScholarWorks, WMU's institutional repository, and have logged over 15,000 downloads so far.

"I think reviewing the decisions that were made, and the information that was released at a given time, could be very valuable for future planning," Orlowska says. "Judging by the number of downloads around the world, someone is clearly analyzing the data. I can't be sure what they're using it for, but it's possible that their scrutiny might help us be better prepared the next time around."

The development of these data sets has been one of several ongoing efforts to collect data and stories for current and future research about the pandemic. Other initiatives include archiving regional government COVID-19 information webpages and dashboards, and a survey focused on collecting personal memories and experiences from southwest Michigan residents during the pandemic.