I'm a data engineer in Chicago. I've spent 15+ years wrangling petroleum and geoscience data — the messy, vendor-locked kind that lives in Petra and GeoGraphix projects and PPDM schemas — and building the pipelines and tooling to set it free. These days that thread runs through modern data engineering stacks (Databricks, Prefect, dbt) and into agentic AI experiments.
- pg_ppdm — the PPDM 3.9 well-data model DDL, converted from Oracle to PostgreSQL. The industry standard schema, minus the Oracle license.
- purr_petra_cli — query S&P Petra projects from the command line and export well-centric JSON for 11 data types. No Petra install required.
- purr_geographix / purr_petra — API-first extraction for GeoGraphix and Petra project data.
- power_puddle — can ComEd's grid handle hyperscaler growth in Northern Illinois? A Prefect + dbt + Grafana pipeline that finds out.
- kingfisher_wells — a geospatial CI/CD experiment with Databricks Asset Bundles: comparing well locations across S&P, Enverus, and the Oklahoma Corporation Commission.
- this_is_fine — agentic wildfire detection and risk assessment fusing NASA FIRMS thermal anomalies, NOAA fire weather, EPA AirNow, and PurpleAir PM2.5 sensors.
- clay-ai-hyperscale — a failed attempt to train the Clay foundation model to detect data centers from satellite imagery. The post-mortem is the point.
- Email: bryan@purr.io
- Web: purr.io — home base, freshly rebuilt and growing
- The cat distribution system also assigned me thequirkykitty.com