mice
mice can finally predict, not just estimate, from multiply imputed data.
A side-by-side editorial comparison of prospectr and robscale — release velocity, themes, recent moves, and the top alternatives to consider.
prospectr spent its biggest release in years fixing spectra it had been quietly mangling.
prospectr provides the signal-processing layer for near-infrared and visible spectroscopy in R — Savitzky-Golay and gap-segment derivatives, standard normal variate, detrending, continuum removal, splice correction, plus calibration sampling algorithms like Kennard-Stone and DUPLEX and readers for ASD and BUCHI NIRCal instrument files. The May release is the substantial one: a long list of corrections to functions that were returning wrong or missing values rather than failing.
A robust-statistics package rewrote its estimators in SIMD C++ and went to CRAN in two weeks.
robscale computes robust scale and location estimators, and its pitch is speed: 21 to 26 times faster than stats::mad, 37 times faster than stats::IQR on small samples, with comparable margins over robustbase for Qn and Sn. The March 2026 releases took it from a GitHub project to a CRAN package carrying eleven estimators, all with confidence intervals.
prospectr provides the signal-processing layer for near-infrared and visible spectroscopy in R — Savitzky-Golay and gap-segment derivatives, standard normal variate, detrending, continuum removal, splice correction, plus calibration sampling algorithms like Kennard-Stone and DUPLEX and readers for ASD and BUCHI NIRCal instrument files. The May release is the substantial one: a long list of corrections to functions that were returning wrong or missing values rather than failing.
The recent work is corrective rather than additive, and several items changed results silently before being caught. continuumRemoval() derived its convex-hull boundary offset from a fixed one-wavelength assumption that broke for fine-resolution spectra or non-nanometre units; cochranTest() passed an invalid argument name to prcomp() and produced incorrect principal component scores; readASD() silently dropped spectra in one branch of its text path. Two file readers were leaking connections. Alongside that runs a smaller thread of decoupling preprocessing steps from each other, most visibly detrend() gaining an snv argument so polynomial detrending can run without the SNV transform that Barnes et al. bundled with it.
The detrend() decoupling is the only recent addition and it fits a broader pipeline-composition direction, so similar separation of other bundled preprocessing steps is the plausible next move. The misspelled substraction argument now carries a deprecation warning, which schedules its removal for a future release.
robscale computes robust scale and location estimators, and its pitch is speed: 21 to 26 times faster than stats::mad, 37 times faster than stats::IQR on small samples, with comparable margins over robustbase for Qn and Sn. The March 2026 releases took it from a GitHub project to a CRAN package carrying eleven estimators, all with confidence intervals.
Three releases in a fortnight walk a clear line: expand the public API, submit to CRAN, then tune. The 0.5.4 work is where that tuning shows, and it is unusually specific about hardware, raising sorting-network thresholds after benchmarking and dropping the AVX-512 path entirely in favour of a shorter AVX2-first dispatch chain. The build-fix lists are long, which is what a package fighting compiler and TBB variation across CRAN's platforms looks like.
The dispatch hierarchy has been simplified once already; further releases most likely continue narrowing the SIMD surface and hardening the configure step rather than adding estimators. A new estimator would be the surprise.
Other Infra & APIs products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either prospectr or robscale.
mice can finally predict, not just estimate, from multiply imputed data.
A market-microstructure toolkit that keeps adding estimators as the papers land.
A vowel-analysis package trimming dependencies after an email address got it archived.
The R half of the EMU speech database system, fixing what was quietly broken.
A Bayesian model-averaging package spending its 2.0 on memory, not methods.
tidyplots keeps rebuilding its own foundations rather than layering around them.
See all prospectr alternatives → · See all robscale alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. prospectr and robscale are shipping at a similar cadence (velocity 0.0 vs 0.0, both within Sparkpulse's "active" band). See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. prospectr and robscale are shipping at a similar cadence (velocity 0.0 vs 0.0, both within Sparkpulse's "active" band). For your specific use case, the alternatives sections above list other Infra & APIs products to evaluate alongside.
Top prospectr alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "prospectr alternatives" section above for the current picks, or visit /alternatives/prospectr for the full list with editorial commentary on each.
Top robscale alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "robscale alternatives" section above for the current picks, or visit /alternatives/robscale for the full list with editorial commentary on each.