A Claude Code skill that makes Claude an expert in the TimeDataModel Python library — the standard for self-describing, bi-temporal time series data in energy and forecasting pipelines.
Once installed, Claude automatically applies this knowledge whenever you work with TimeSeries or TimeSeriesTable objects. No commands to remember — it just works.
- DataShape selection — when to use
SIMPLE,VERSIONED,CORRECTED, orAUDITbased on your modeling needs - Creating time series — from pandas, Polars, NumPy, and PyArrow, with the right
Frequency,DataType, and metadata - Bi-temporal modeling — structuring forecasts with
knowledge_time+valid_timefor full auditability - Enums — complete
Frequency,DataTypehierarchy, andTimeSeriesTypereference - Workflows — import/export, slicing, unit conversion (pint), geospatial filtering, and validation
Search for it in Claude Code:
/find-skills timedatamodel
Or install directly:
/plugin install timedatamodel
After installing, just describe what you want:
"Create a versioned TimeSeries of hourly wind power forecasts from this pandas DataFrame"
Claude will produce correct, idiomatic TimeDataModel code — right DataShape, right enums, UTC-aware timestamps — without you needing to look anything up.
TimeDataModel is an open-source Python library by Rebase Energy for modeling time series data with rich, machine-readable metadata. It is designed for energy systems, forecasting, and any domain where data provenance and bi-temporality matter.
pip install timedatamodel