ASA DataFest at UCLA · May 2026
A dataset, five people, 48 hours.Healthcare Data Analysis
I coordinated a five-person team working with de-identified patient data at ASA DataFest.
- Role
- Team coordinator
- Team
- Five people
- Tools
- R, ggplot2, Spearman correlation
48 hours, five people
At ASA DataFest at UCLA, I coordinated a five-person team working on de-identified patient data. My work centered on stress correlations and on the presentation connecting our four analyses: geographic access, stress and social needs, care continuity, and patient-portal engagement.
From records to patients
I joined survey, demographic, and encounter tables in R, reconciled identifiers and types, parsed dates, and aggregated repeated questionnaires before patient-level tests.
Features included ordinal stress scores, encounter counts, distinct visit days, and gaps between visit dates. Keeping question responses, encounters, and patients separate as units of analysis was important.
Statistical analysis
I used Spearman rank correlation with Benjamini–Hochberg correction and nonparametric cohort comparisons. The work also included age-bin-adjusted logistic models exploring screening and social needs by patient-portal status.
These were exploratory associations in observational records, not causal effects or validated clinical predictions. The underlying data and unpublished patient-level results are not shared here.
Making the work legible
I built ggplot2/R Markdown reports and scripted a five-minute patient-journey presentation synthesizing the team’s analyses. The coordination and communication work was part of the project, alongside the statistical pipeline.
