Dataset Playbook to Help Canadians Understand COVID Trends
August 16/2021
By Françoise Makanda, Communications Officer at DLSPH
A popular student-led pandemic dashboard just published its ‘playbook’ allowing anyone to use its COVID data.
“We have some graphs and maps showing how the data can be used. It’s very illustrative, but it allows people to get an idea of what you can do,” says PhD Candidate Isha Berry, co-lead creator of the dashboard, COVID-19 in Canada. The methods paper was featured in The Scientific Data journal.
The paper focuses on data descriptors that allow researchers to understand how data is curated for daily reporting. While it gives researchers context, it also makes the data accessible for immediate use.
“The data descriptor is just an overview of our whole methodology process. We have a technical report that we have written up, but we knew that it’s not super easy to always find a technical report or to know what to look for,” says the epidemiology student.
The paper shows how the students worked with COVID data that’s reported in differing ways. For instance, provinces like British Columbia do not report cases on weekends. Some provinces offer COVID case breakdowns by age, some don’t.
Berry wants more information on travel-related cases since borders are opening again and were initially the virus transmission source. “As we start opening it will be really useful to continue to gather that data because I think [border transmission] will be a bigger source at some point if we are able to contain our domestic transmission,” says Berry.
Dashboard co-lead and PhD student Jean-Paul Soucy wants a breakdown of case severity by vaccination status.
“I hope we see more regional data on hospitalizations and data on severe outcomes by vaccination status. The latter will help bolster the case for vaccination. Hospitalizations due to COVID are in the spotlight in the United States right now, and the Department of Health and Human Services provides information on how local hospitals are affected by the pandemic to a level we can only dream of in Canada.”
Berry has learned that observing a weekly case average—as opposed to daily case counts—provides better insights into trends. It helps her team see how things will unfold before they happen. But, like any open-access software, users have free rein over how they interpret the data. Transparency is key.
“It really allows people to conduct analysis in real-time, especially when this is open data that’s published every day,” she says. “We really rely on lots of these open, publicly available data sets to forecast, understand and anticipate what policies might need to be implemented.”
PhD Program Director and Prof. Laura Rosella share a similar sentiment. Open access data is important for transparency and reproducibility.
“Of course, not everyone will have the ability to use the data but knowing it’s there and that people with different skills and abilities can work on it helps build trust and can enhance the insights by opening data analysis up to a broader range of skillset. Both open and closed data can be used the wrong way, so I don’t see any greater risk and the benefits of the transparency and greater insights far outweigh that risk.”
Rosella is proud of students who have taken up the project one year and a half ago. “One reason I think they were able to push this so far is the willingness of this generation of students to see difficult problems as a team sport and one where putting different skills together can achieve something really powerful.”
At that time, it was just Berry and Soucy, hunched over their computers. The team grew to 16 U of T and University of Guelph students, and by June 2021, the team had automated some processes to make data curation easier.
“When we started, there was no one else really doing it. It didn’t really exist for Canada, and now there’s so many different data sources that have been so open,” Berry says. “I don’t know to what extent we’ve been helping that, but I think something that’s really been cool is to see that.”
This work was made possible in part thanks to U of T’s student engagement award.