I am a computational social scientist and Assistant Professor of Sociology (and, by courtesy, Public Health) at Purdue University. My research examines how cultural meanings structure health, illness, and social categorization, with particular attention to gender and disease. I specialize in computational approaches to text (e.g., word embeddings, transformer-based language models, and topic modeling) and draw on data such as social media, news, electronic health records, and administrative records on violent death. Across my current projects, I seek to address a central question: How do cultural constructions of disease and identity impede or facilitate well‑being?
My work also circles around a core methodological challenge: How can scientists measure semantic information in text data? Advancing computational text analysis requires both technical and theoretical innovation. As my research highlights, these methods require scientists to mathematically model the nuanced ways in which human language encodes and conveys meaning.
After earning my Ph.D. in Sociology from UCLA in 2022, I completed a National Library of Medicine–funded postdoctoral fellowship in biomedical informatics at UC San Diego (2022–2024). This training allowed me to bridge sociological perspectives on health and culture with computational approaches and problems central to biomedical informatics. Prior to graduate school at UCLA, I conducted research on social media and health at Seattle Children’s Hospital and Research Institute. I received my B.A. in Sociology from the University of Washington in 2014, and I am a Seattle native.
Beyond my research, I am deeply invested in computational and quantitative social science pedagogy and in strengthening the computational social science community. I co-organized and taught in the Summer Institutes in Computational Social Science (SICSS) at UCLA for three years, helping to mentor early‑career researchers across disciplines. At Purdue, I regularly teach Introduction to Statistics in Sociology (SOC 382). My teaching emphasizes both technical competence and reflexivity. I push students to understand not just how methods work, but how our analytic and data choices shape the knowledge we create.
Contact me: aaashelm at purdue.edu
My work with Andrei Boutyline "Meaning in Hyperspace: Word Embeddings as Tools for Cultural Measurement" is in published in Annual Review of Sociology.
My article with Rachel Best, titled, "Disease Frames and Their Consequences for Stigma and Medical Research Funds" is out in Social Science and Medicine.
My newest research on obesity diagnosis with a team at UC San Diego, "Leveraging Diagnosis and Biometric Data from the All of Us Project to Uncover Disparities in Obesity Diagnosis" is out in Obesity Pillars.