Data Scientist (Primary Care Analytics, Health Services Research And Development)
Current• Co-authored a study on 1.3 million high-risk patients to assess engagement with primary care telehealth during the COVID-19 pandemic. Collected, cleaned, and processed cohort data using SQL and R. Developed the modeling script for a multinomial logistic regression in R and contributed to the interpretation and analysis of the results. • Co-authored a study analyzing disparities in home health usage between rural and urban patients (2019-2021). Built the patient cohort in SQL, utilized R to create a hurdle model (logistic regression and negative binomial regression), and provided comprehensive interpretation of the study's findings. • Built data architecture in SQL to extract, transform, and load over 1 billion activity transactions from VA computer systems, covering 7,000+ primary care physicians. This data, captured at the millisecond level, was used to calculate the duration of provider activities (e.g., secured messages, chart reviews, outpatient orders). Trained Relative Weight Units at multiple levels to improve performance through simultaneous events processing and state optimization. • Trained a Hierarchical Agglomerative Clustering model in Python to categorize VA facilities into subclusters based on various provider activities, enabling targeted analysis of facility performance and identifying potential improvements.