Arimancy

About

Arimancy is a data products company. We build model-ready time series datasets for power-market forecasting — cleaned, joined, and validated before they reach your notebook. ERCOT hourly demand and weather is the first product; CAISO, PJM, MISO, and NYISO extend the same schema next. Nothing else is named or promised until it ships.

How we build

  1. 01UTC-first joinsEvery join happens in UTC before local time zones are applied. Power markets observe daylight saving on their own local clock, and a join performed in local time silently duplicates or drops an hour twice a year. Arimancy resolves every timestamp to UTC first and applies time zone conversion only for display, so spring-forward and fall-back transitions never show up as gaps or repeats in the demand series.
  2. 02Verified sourcesEvery input is checked against its original publisher before it enters a dataset, not just the aggregator that redistributes it. Source and license are recorded per column, so anyone using the data can trace a number back to the agency or instrument that produced it. If a source can't be verified, it doesn't ship.
  3. 03Documented caveatsReal-world time series have gaps: sensor downtime, delayed reporting, retroactive revisions. Rather than smoothing these over or burying them in a README, Arimancy flags every imputed or estimated value directly in the data, row by row, so downstream models can account for uncertainty instead of inheriting it silently.
  4. 04Permissive licensingEvery dataset ships under CC BY 4.0. Attribution is the only condition — no seat limits, no commercial-use carve-outs, no separate enterprise tier. A forecasting team should be able to use the data the day they find it.

Founder

Arimancy is built by a software architect with more than a decade in data platforms, working from Oregon. His academic background spans nuclear engineering, physics, and astronomy — fields where a mismeasured input compounds fast. He started Arimancy to apply that standard to power-market data: verified sources, documented caveats, and joins that hold up under scrutiny. The company is independent, building one dataset at a time.