Compare ten stable workflows¶

The ten workflow orchestration routes were published with 1.2.0 as module-scoped APIs. A workflow organizes scientific decisions and records; it is not a guarantee of calibrated statistical inference. Select a scientific question to narrow the table. The entire table remains readable with JavaScript disabled.

Research need

Search method

10 of 10 entries shown

Stable entry point	Observed data and estimand	Required reporting / boundary
`run_sparse_fpca_workflow()`	One sparse functional coordinate; mean/covariance/modes	Native timestamps, smoothing, noise, PSD and score failures
`run_sparse_mfpca_workflow()`	Paired sparse x/y, joint covariance	Shared observed times and cross-channel covariance; not two separate scalar fits
`run_sparse_mfpca_async_workflow()`	Coordinate-asynchronous x/y times	Separate time support per channel, never invented pairing
`run_sparse_multilevel_workflow()`	Sparse participant/trial functions	Cluster IDs, decomposition and independent-unit support
`run_sparse_prediction_workflow()`	Prediction from partially observed trajectories	Train/calibration separation, horizon and coverage scope
`run_fpca_workflow()`	Common-grid continuous trajectories	Coordinate scaling, rank selection, fit support
`run_function_on_scalar_workflow()`	Functional effect of scalar covariates	Design matrix, independence, inferential target
`run_functional_mixed_effects_workflow()`	Repeated-trial functional effects	Random effects, covariance, cluster resampling; W3 limitation #289
`run_generalized_functional_workflow()`	Binary / grouped-binomial / count functions	Family/link, denominators, exposure and hierarchy
`run_recurrence_workflow()`	Recurring states and nonlinear structure	Embedding, delay, radius, Theiler exclusion and RQA limits

The original ten-workflow article gives the exact scientific decision sequence, and the product qualification report gives the engineering evidence. This table does not supersede documented preconditions, exclusions or scientific refusal behavior.