Mining for Meaning in COVID-19

A professor uses the pandemic as an opportunity to teach analytics students how they can use data to make a difference.

ANALYTICS IS HAVING a major moment, as researchers mine the deluge of data being produced during the pandemic. Educators, too, view this time as an opportunity for students to discover how analytics can provide real-time insights into a crisis.

Among them is Kathleen Iacocca, a professor at the Villanova School of Business in Pennsylvania. This past spring, she already had a text analysis included on the syllabus for her advanced analytics course. When the time came, it made perfect sense to focus that assignment on COVID-19. She asked students to determine ways that companies could use text mining to identify patterns coming out of the pandemic.

Student teams primarily used the R programming language to scrape data from the web—including posts on Twitter as well as content from major news outlets such as NBC, BBC, and CNN. Students prepared the data using Tableau Prep, and then used JMP data analysis software to explore the text.

Student teams determined several ways companies in different sectors could benefit from text mining. For example, teams pointed out that health insurance companies could text mine social media posts to better understand how social distancing affected people’s mental states; companies then could use that knowledge to tailor support services to their customers. Others realized that healthcare providers could use text mining to scan medical transcripts to discover whether preexisting conditions correlate to more serious cases of COVID-19 or even predict which patients are more likely to be admitted to the hospital. News outlets could mine social media posts to identify whether Americans were most concerned with identifying early symptoms, reducing death rates, or reviving the economy.

Iacocca included an essay question on the final exam asking what actionable insight on COVID-19 stakeholders could gain from using text data. Some students analyzed news stories for insights on the economy, while others discussed the potential to use text mining on physicians’ notes to better understand recovery trends.

This assignment showed students how they could use data analytics to assist in crisis management and recovery, says Iacocca. “While text mining analysis is not a traditional way of thinking about contributing to the pandemic, I wanted to get the point across that everyone has a skill that can be used to improve the human condition,” she says. “In future semesters, I will bring in other topics that revolve around improving the human condition, whether those topics involve social justice issues, health issues, or something else. Students enjoyed feeling like they could use text mining for positive change.”