Justin MpangaPublic health · Statistical modeling
CDC Diabetes Burden: Trends & Model Benchmark
In this CDC snapshot, the age-adjusted estimate shifts from 6.5% in 2015 to 6.6% in 2024. The linked notebook includes subgroup intervals and a 2020–24 model holdout.
Data analyst · Healthcare + research
Turning complex data into clear decisions.
I build thoughtful data workflows, predictive models, and visual stories that help biomedical teams move from questions to measurable outcomes.
05+Years in analytics
03Core disciplines
360°Workflow ownership
1Curious mind
01 / Selected work
Analysis with
an outcome.
A few ways I have used data to make research more precise, visible, and actionable.
Healthcare operations · Reimbursement
Uninsured-Care Reimbursement in DuckDB SQL
Texas leads this snapshot at $3.16B; California follows at $2.48B, mostly from testing. The linked notebook includes the audit, state mix, and provider rankings.
Infectious disease · Data engineering
Lyme Surveillance: Multi-Era ETL
The R pipeline combines 893,519 CDC records across three surveillance eras, labels each row by period, and checks schemas before appending.
Data storytelling · Visualization
Uninsured Care Reimbursement and Diabetes Burden
The dashboard specification keeps provider payments separate from population estimates and calls for visible uncertainty, source dates, and measure definitions.
02 / The approach
Curious by nature.
Rigorous by habit.
I’m Justin, a data analyst specializing in clinical research and healthcare analytics. I enjoy finding the useful signal inside complicated datasets, then making that signal understandable to the people who need it.
My work spans the full journey from ETL and feature engineering to predictive modeling and visualization. The best analysis, to me, is technically sound, easy to interrogate, and connected to a decision.
Build
Python / SQL / Pandas
scikit-learn / PyTorch
ETL / Feature engineering
Communicate
Tableau / Power BI
Plotly / Matplotlib
Clear, useful storytelling
03 / Start a conversation
Have a meaningful data problem?
Tell me what you are working on. I’m always interested in thoughtful questions and the people asking them.