01 / STATISTICAL MODELING
Cognitive recovery after long-haul flight
Explored time to recovery through survival analysis, connecting performance measurements with sleep and trip context.
BIOSTATISTICS / DATA SCIENCE / RESEARCH SOFTWARE
Clarity in the complexity.
I turn research data into analysis people can inspect, understand, and trust.
From joining records and defining outcomes to checking models and communicating limits, my work makes the reasoning visible.
Evidence-led. Reproducible. Human-readable.
01 / DATAStructure the signalClean inputs · clear definitions
02 / MODELTest the questionMethods · assumptions · fit
03 / VERIFYMake it defensibleDiagnostics · interpretation
FROM QUESTION TO EVIDENCE
01 / 04SCROLL TO EXPLORE
Statistical modelingResearch data workflowsInteractive analysisValidation & reporting
01 / SELECTED WORK
Work worth looking into.
Six snapshots of the questions and capabilities behind my research work. Project materials and participant-level data remain private.
02 / DATA ENGINEERING
Sleep data integration
Aligned wearable, sleep log, and flight records into analysis-ready features across trip and time boundaries.
03 / INTERACTIVE ANALYSIS
Research dashboards
Built exploratory views for researchers to compare routes, crew roles, fatigue, and cognitive performance.
04 / RULES & VALIDATION
Flight-duty rule validation
Translated complex screening rules into traceable outputs and checked results against source records.
05 / STUDY DESIGN
Study power & simulation
Evaluated study designs using endpoint-specific assumptions, power calculations, and repeated-measures simulation.
06 / AUTOMATION
Reporting automation
Made recurring research metrics more consistent with structured completeness rules and reviewable outputs.
These are concise public summaries. Detailed methods, research files, and participant data are not published here.
02 / THE APPROACH
Good analysis is a chain of decisions.
I make the important links inspectable, from first input to final interpretation.
Define clearly
Establish the question, unit of analysis, and eligibility rules before counting or modeling.
Build carefully
Align records, timestamps, and context in a reproducible workflow.
Model appropriately
Choose methods for the question and data, then state the assumptions behind them.
Check honestly
Review diagnostics, missingness, and source-to-output agreement before sharing conclusions.
03 / CAPABILITIES
Tools in service of better questions.
A mix of statistical thinking, practical coding, and clear communication.
Statistical modeling
Survival analysis, Cox regression, mixed-effects models, repeated measures, simulation, and diagnostics.
Data workflows
R, Python, SQL, tidyverse, data validation, feature engineering, and reproducible reports.
Research interfaces
R Markdown, Quarto, ggplot2, Shiny, Plotly, and Excel reporting.
SELECTED TRAINING
Survival Analysis in R SQL for Data Science Intermediate Python Joining Data with pandas
04 / LET’S CONNECT
Let’s make the complex clear.
Interested in biostatistics, research analytics, or the thinking behind these projects? I’d love to talk.