Research Activities
At the Johns Hopkins Bloomberg School of Public Health, I contributed to Dr. Ana Rule's research lab, where our work addressed public health concerns related to e-cigarette use. My focus was on investigating toxic metal exposure by analyzing e-cigarette aerosols and human biospecimens, while also studying user behaviors and device characteristics to understand the full scope of exposure.
Concurrently, I was a member of Dr. Ahmed Hassoon's research team, where we investigated the potential for diagnostic errors by Large Language Models (LLMs) across various clinical scenarios. Our goal was to advance AI safety in medical settings, ensuring that these powerful tools were applied both effectively and responsibly to improve patient outcomes and benefit the general public.
During my undergraduate studies, I worked in Dr. Hsing Jasmine Chao's lab at Taipei Medical University, focusing on air pollution and children health. I gained extensive hands-on experience in exposure assessment, environmental sampling (especially air sampling), and spatial analysis, which initiated my interest in the role of the environment in human health and society. Most importantly, her mentorship introduced me to the world of science and research.
Publications
Coming soon...
Media
Can your AI save a life when the doctor's makes diagnostic errors? (August 2025). LinkedIn coverage of Johns Hopkins research testing Large Language Models' ability to course-correct physicians' diagnostic errors during initial patient presentations. View post
Can AI Catch What Clinicians Miss? (June 30, 2025). Published on PHAISE AI Blog, exploring the potential of artificial intelligence in reducing diagnostic errors and improving clinical decision-making. Read the full article
Student Spotlight: Xiaoyi (Showry) Peng (February 20, 2025). Featured in Johns Hopkins Bloomberg School of Public Health student spotlight, highlighting my journey and research interests in public health. Read the full article
Presentations
Peng, X., Feder, K.A., Rule, A.M., & Aherrera, A.D. (2025, March). Characterization of Metal and Nicotine Exposure from Electronic Cigarette Use (Vaping) via Urine Biomarkers. Poster presented at the 31st Annual Meeting for the Society for Research on Nicotine and Tobacco, New Orleans, LA.
Peng, X., Irimia, R., Leo, H., Bandeira, A., Li, L. A., & Hassoon, A. (2024, November). AI-Powered Diagnosis: A Comparative Study of Large Language Models in Reducing Diagnostic Errors. Poster presented at the Public Health + AI Strategic Endeavors (PHAISE) AI Research Day, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD.
Peng, X. & Zhang, B. (2022, October). A Community Assessment and Chronic Disease Prevention Survey in Longtian Borough, Songshan District, Taipei City. Poster presented at the Taiwan Public Health Association Annual Conference, Taipei, Taiwan.
Academic Projects
Johns Hopkins University
2023 - 2024Statistical Machine Learning: Methods, Theory, and Applications
August 2024 - October 2024 Instructor: Vadim Zipunnikov, PhDPredictive Modeling of Long-term Mortality Risk Using NHANES Data
For this project, I developed and systematically evaluated a suite of machine learning models—including Random Forest, Gradient Boosting, SVM, and regularized logistic regression—to predict 9-year mortality from NHANES data. I applied rigorous feature selection to identify 38 key predictors and implemented a Random Forest-based imputation strategy to handle missing data. To determine the optimal model for potential clinical application, I performed a comparative analysis based on ROC curves, sensitivity, and specificity.
Spatial Analysis IV: Spatial Design and Application
March 2024 - May 2024 Mentor: Frank C. Curriero, PhDAssociation Between Poultry Farm Proximity and Agricultural Air Pollution in South Delaware
In this project, I designed and implemented a spatial analysis workflow in R to investigate agriculture-related air pollution. I mapped ammonia (NH₃) concentrations from mobile monitoring data, integrating both spatial and temporal coordinates within census tracts. To model exposure, I quantified poultry farm density using a three-tier buffer system and then developed a GLS-based multiple linear regression model to assess the statistical relationship between NH₃ levels and proximity to these farms.
Data Science for Public Health II
March 2024 - May 2024 Instructors: Brian Caffo, PhD; Ahmed Hassoon, MDLiterature Review HelperGPT
To streamline the systematic literature review process, I developed "Literature Review HelperGPT," a full-stack web application using Flask and a custom HTML/CSS frontend. The application integrates Large Language Models (LLMs) to enable batch processing of scientific articles through an automated workflow that includes document querying and CSV data export. I focused on designing an intuitive user interface to simplify article management and the consolidation of research findings.
Applied Environmental Health Practice
October 2023 - December 2023 Instructor: Megan Latshaw, PhDSociocultural Public Health Impact of Broadband Internet Access in Tasso Island, Sierra Leone
As part of a collaborative project with a consulting firm, I evaluated the public health, sociocultural, and economic impacts of broadband internet implementation on Tasso Island, Sierra Leone. I led the stakeholder engagement process, conducting structured interviews with local residents to assess the implications of this new connectivity. My analysis centered on identifying economic development opportunities, and the project culminated in my development of policy recommendations for sustainable broadband integration.
Taipei Medical University
2021 - 2022Spatial Epidemiology & Applications of GIS
December 2021 - January 2022 Instructor: Ting-Wu Chuang, PhDImpact of Epidemic Alert Level 3 on PM2.5 Levelsin Taipei City
This project involved analyzing PM2.5 trends across 12 districts in Taipei City during the COVID-19 lockdown. Using data from seven monitoring stations, I applied the Ordinary Kriging geostatistical method in ArcGIS to generate monthly concentration maps and calculate zonal statistics. My subsequent statistical analysis confirmed a significant 25–40% reduction in PM2.5 levels in seven districts, providing quantitative evidence for use in environmental policy and urban planning.
Teaching & Leadership
Statistical Methods in Public Health I (140.621), Johns Hopkins Bloomberg School of Public Health
Teaching Assistant | Instructors: Karen Bandeen-Roche, PhD & Marie Diener-West, PhD | 2024
Laboratory for Environmental health II, Taipei Medical University, Taipei, Taiwan
Teaching Assistant | Instructor: Hsing Jasmine Chao, ScD | 2023
Taiwanese Student Association, Johns Hopkins University, Baltimore, MD
Medical Campus Representative | 2024-2025
Epidemiological Investigation Team, Taipei Medical University, Taipei, Taiwan
Captain | 2021-2022
Executive Secretary | 2020-2021
Environmental Health Group Member | 2019-2020
Technical Skills
Programming: Python, R, SAS, SQL, Stata, Linux, Flutter, Firebase, React
GIS & Data Management Platforms: ArcGIS, REDCap, QGIS, GeoDa
Creative Softwares: Adobe Creative Suite (Photoshop, Lightroom), Final Cut Pro, Blender
Languages: English, Mandarin Chinese