About

Cameron Grulke

I designed and built Dunmore Research end to end: the data pipeline, the 127,457-entity graph, the public API, and this site.

This page covers what that involved, for anyone curious about the engineering behind it.

Email LinkedIn
API endpoints
20
Entities indexed
127,457
Data tables integrated
14
Public data sources
5

What building this involved

Data engineering

Wrote the ingestion pipeline that pulls FBI Crime Data Explorer, FEC bulk filings, Census government finance & ACS demographics, USAspending grant awards, and Federal Register activity into one unified, deduplicated graph of every US court, police department, jail, CPS office, and federal agency.

Data validation

Root-caused a systemic parsing bug in a live FBI API integration that had silently corrupted 555,000+ crime records. Rather than ship it, I removed the feature, documented the cause, and rebuilt the analytics layer on data that held up under scrutiny.

Full-stack delivery

Node/Express + SQLite API with tiered rate limiting and a self-serve API-key flow, auto-generated interactive docs, and a from-scratch vanilla JS/CSS frontend, including a custom MapLibre heatmap-to-point rendering pipeline for 124,000+ geocoded markers that stays smooth at every zoom level.

Iteration

Revised the map's interaction model based on usability feedback: fewer clicks to reach an individual entity, a density heatmap in place of cluster counts. Shipped the redesign the same day.

Get in touch

Questions about the data, the architecture, or anything else in this project are welcome.

camerongrulke@gmail.com LinkedIn →