hello@vectoranalytics.org
Data systems for public‑health programs.
I build analytical applications and dashboards. Working software that replaces fragile manual processes with something a program can actually rely on.
- 10+ yrs Clinical research and FDA-adjacent data environments
- JHSPH Graduate training, Johns Hopkins Bloomberg School of Public Health
- Full-stack Ingestion, statistics, interface, deployment, all built and shipped together
- One person The same person who builds the system picks up when it breaks
Made to answer the next question, not just the current one.
Public health programs generate a lot of data. Most manage it with workflows that only one person fully understands, that take a week to produce outputs a funder asked for in an email, and that break quietly when something upstream changes.
I build systems underneath that. Data ingestion, analytical rigor, the dashboards on top, all running and monitored. The stats are done by someone with the training to do them right and the experience to put them into production.
Ingestion & pipelines
Sentinel feeds, line lists, REDCap exports, lab CSVs. Validated, versioned, and dated so you can defend the numbers later.
Statistical core
Methods chosen for the data structure and the question, not for what the charting library can render. Standard errors that respect the design. Estimates validated against published reference numbers and surfaced with their confidence intervals rather than tucked behind the chart.
Dashboards & maps
Choropleths, time series, filterable tables, downloadable views. Used directly by program staff for the deliverables that have to ship on schedule.
Deployment & handoff
Hosted, monitored, documented. Survives staff turnover, scope changes, and audit cycles. That is the job.
Built like infrastructure. Supported like it too.
-
01
One person, beginning to end.
The person scoping the work is the person building it, and that same person picks up when something breaks. I'm accountable to the system from start to finish.
-
02
Statistics done correctly.
Methods chosen for the question, documented with assumptions and limitations, validated against published estimates where they exist, and surfaced in the UI rather than buried in an appendix.
-
03
Engineered for production.
Tested, version-controlled, deployed, monitored. A working system on Monday morning, not a notebook your team can't run when the analyst is on leave.
-
04
Built to outlast the engagement.
Documentation a new analyst can use on day one. Runbooks for the parts that page someone. A handoff that does not require pinging the previous person six months later.
One name on the contract. The same name on the support call.
Vector Analytics is a solo public-health data systems practice. I work with state and local health departments, academic research programs, and global health organizations on surveillance platforms, analytical pipelines, and spatial methods.
I'm Nate, a master's-trained biostatistician with current graduate work at the Johns Hopkins Bloomberg School of Public Health. My last decade has been spent in clinical research and FDA-adjacent data environments, where data pipelines and analytics have to survive a line-by-line audit. That is the standard I hold these systems to.