Things I built for myself and decided to share: a synthetic distribution dataset, a growing ML playground, a handful of reliability calculators, and the methodology papers that back the writing. MIT-licensed, forkable, and yours to keep.
Realistic synthetic data you can load and run analysis against without a CEII form.
Open synthetic distribution utility. 238,104 customers modeled across 23 feeders, two substations, and a small generation fleet. OpenDSS-ready, pandas-friendly, nightly regenerated. Hosting-capacity studies, outage-pattern analysis, and ML training without the NDA.
Generator scripts, dataset schema, and the dbt project that assembles it. Fork, run, modify. If you build something interesting, tell me.
Working systems I built on top of the datasets. Apache-2.0, clone-and-run. The architecture patterns come straight out of the articles; the code is proof they ship.
Twenty-three notebook-style guides for the ML techniques that actually matter in utility operations. Each one is a working model with real inputs, real outputs, and a plain-language methodology.
Outage prediction, load forecasting, hosting capacity, predictive maintenance, FLISR optimization, volt/var, DER scenario planning, anomaly detection, rotating-equipment health, OpenDSS integration, loss analysis. Fundamentals through advanced, all runnable.
Classification model on SP&L fault records. Feature engineering, model selection, and the pieces that matter for deployment at a real utility, including how to hand the model off so an engineer can retrain it next quarter.
Reliability math in the browser. No signup, no email wall. Paste your SAIDI, get your burn rate.
Convert a SAIDI or SAIFI target into a per-feeder error budget. Derive daily burn rate, per-event opportunity cost, and the threshold at which you should trigger a reliability freeze.
Price an outage in dollars, not customer-minutes. Layered by customer class, duration, time-of-day, and regulatory penalty exposure. Useful for CAPEX prioritization when the numbers have to go in front of a CFO.
The papers and frameworks that sit behind the tools, in readable form.
The framework page. What an error budget means for a distribution circuit, how to derive it from an IEEE 1366 report, and what operational practices sit on top of it (burn-rate alerting, reliability freeze, blameless postmortem).
Seven-part series porting Google's Site Reliability Engineering methodology onto distribution reliability. Budget, don't scorecard.
Things I built to learn something or get better at it. Browser-playable, no install. If a game can teach the material faster than a slide deck can, I want to know.