I build data systems for health programs.

The whole analytical chain, from first data to final report, including transparent data ingestion, estimates with uncertainty, and interactive dashboards, maps, and reports your program can rely on. I'll build it, deploy it, and hand over the keys.

Email me

Why my work holds up

  • 01

    Regulated data

    where the numbers have consequences

    My work has taken me from pharmaceutical research and manufacturing to Medicaid quality measurement and beyond. Many settings, always with the same requirements: the work has to be traceable, reproducible, and defensible when an auditor with a grudge comes digging.

  • 02

    Systems

    for this year, next year, and the next

    I build data systems with the next data cycle and the next person in mind. I document how they work, what happens when something changes, and how to recover when something breaks. A production system has to survive more than its launch.

  • 03

    Scrutiny

    for the person who wasn't in the room

    Every analysis eventually lands in front of someone who didn't build it. A reviewer, program manager, auditor, or future analyst should be able to understand where the number came from, what assumptions produced it, and what the evidence actually supports.

  • 04

    Reproducibility

    including the parts that disagree

    I have worked with both legacy and modern analytical systems, including parallel SAS and reproducible implementations running against the same data. When they disagree, I trace the difference to its source and write down why. Picking whichever answer looks right is how errors fester.

Everything between the data and the decision

Between a lab export and a progress report there is usually a person, a spreadsheet, and a long afternoon. That works until the person is on leave or the vendor changes a column, and then the number moves for a reason no one can explain.

  • Data that arrives, gets validated, and leaves a record of everything that touched it
  • Estimates that carry their uncertainty, especially when the numbers are small
  • Maps and dashboards built for people who need to use the answers
  • Reports generated from the same data that produced the analyses and visualizations
  • Systems that keep running long after the handover

Let's solve the hard part.

A short description of your data and your goal is enough to start. I'll let you know if it's work I can do.

Email me hello@vectoranalytics.org