masque: Structurally Faithful Development Surrogates for Tabular Data

masque turns a single tabular dataset into a structurally faithful synthetic clone suitable for pipeline development. Experimental-design columns and the NA pattern are preserved exactly; treatment and categorical-covariate level vocabularies are optionally aliased; outcome and numeric-covariate values are re-simulated via a Gaussian copula that preserves the global covariance structure. A private recipe object round-trips a pipeline written against the synthetic clone onto the original data.

Honest claim

masque is not a differential-privacy or anonymisation tool. Its outputs are development surrogates: structurally faithful enough that pipeline code runs unchanged, controlled enough that raw values are not exposed in collaborate mode. See vignette("confidentiality", package = "masque") for the full threat model and limitations.

Two modes

• mode = "local": owner's realistic dev surrogate; preserves names and level vocabularies; may emit observed values; not for external sharing.

• mode = "collaborate": opaque aliasing for treatment and categorical-covariate levels; within-resolution jitter on numerics; integers stochastically rounded; audit_mask() auto-runs at construction.

See also

Useful links:

• https://github.com/max578/masque

• Report bugs at https://github.com/max578/masque/issues

Author

Maintainer: Max Moldovan max.moldovan@gmail.com (ORCID) (Adelaide University)