Researcher Spotlight: Jiyeong Kim
Our Researcher Spotlight Series profiles DataLab affiliated faculty, staff, graduate students and postdoctoral scholars across disciplines to demonstrate the diversity of our community and highlight their amazing work.
Jiyeong Kim is a current PhD student in Public Health Sciences and an inaugural Translational Health Data Science (THDS) fellow. Her project is titled “Helping Caregivers find Tailored Mental Health Information by Applying Natural Language Processing to Online Caregiver Forums.”
Jiyeong’s research addresses mental health issues affecting vulnerable populations—cancer survivors, low-income communities, and most recently, caregivers—and she is all too familiar with how difficult it can be to accurately and efficiently identify concerns in these populations. Typically, if researchers want to learn about the concerns of caregivers—what kind of help and support each caregiver needs—they would need to organize a qualitative interview, a process Jiyeong emphasized requires substantial time and resources. “Even then, these types of qualitative assessments can only interview 10-20 people,” Jiyeong said; and further complicating the research, “it is not easy to find caregivers who are willing to do an interview, and the result of that interview does not represent the needs of all caregivers.”
I am interested in using these technologies to help people…
Jiyeong wants to eliminate these barriers to identifying caregivers’ needs. In her THDS fellowship project, Jiyeong tested the ability of a natural language process approach to identify mental health issues of caregivers of Alzheimer’s disease using an online caregiver forum. Jiyeong collaborated with computer scientists—DataLab’s Dr. Vladimir Filkov and four undergraduate computer science students—to build a data pipeline that scraped tens of thousands of caregivers’ posts and used topic modeling to detect word and phrase patterns within the text. The collaboration was particularly exciting for Jiyeong, because she saw the research as “a bridge between computer science and health science.”
Jiyeong’s project simultaneously and accurately assessed the needs of thousands of caregivers, providing evidence that a natural language processing approach can be an effective tool to identify mental health issues. “With our results,” Jiyeong said, “we can discuss further steps with our connections at the Alzheimer’s Association on how we can use these data sets to support caregivers.” It is not hard to see how computational approaches like Jiyeong’s—though not exhaustive—have the exciting potential to improve health providers’ and public health officials’ ability to detect and rapidly respond to emerging health crises.
Moving forward, Jiyeong plans to use social media data and topic modeling to quickly assess public health issues and their relation to social media. “I am interested in using these technologies to help people,” Jiyeong says, and she wants “to collaborate with clinicians so they can actively make targeted interventions.” As a mother of two young children, Jiyeong is especially interested in addressing mental health issues in mothers and children, specifically using her skillset of bridging computer and health sciences to help address postpartum depression in young mothers.
You can read more about Jiyeong’s research in her recently published paper on caregiver HPV awareness.