PurpleAir
The R client for PurpleAir sensors keeps finding its time-averaging was wrong.
A side-by-side editorial comparison of lstar and rmediation — release velocity, themes, recent moves, and the top alternatives to consider.
A single-cell data store commits to Zarr v3 and range-readable hosting across four language surfaces
lstar stores single-cell data behind one C++ core with Python, R and JS/WASM bindings, and ships a browser viewer that reads the store directly. Zarr v3 is now the default on-disk format across all four surfaces, with zstd compression and sharding that packs many chunks into fewer objects. Viewer stores are compressed per field and resolved at chunk granularity, so a hosted viewer fetches only what it displays. The tag stream carries both lstar and lstar-sc releases.
RMediation shipped a three-normal CDF, then found it was silently wrong.
RMediation is a long-standing CRAN package for confidence intervals on mediated effects, now built on an S7 class hierarchy. Over eight weeks it added ProductNormal3 for serial indirect effects of the form a1*a2*b, folded the engine into the existing pprodnormal naming family, and then replaced that engine outright after finding it returned wrong probabilities without warning. The dev branch is at 1.7.0; CRAN still serves 1.6.1.
lstar stores single-cell data behind one C++ core with Python, R and JS/WASM bindings, and ships a browser viewer that reads the store directly. Zarr v3 is now the default on-disk format across all four surfaces, with zstd compression and sharding that packs many chunks into fewer objects. Viewer stores are compressed per field and resolved at chunk granularity, so a hosted viewer fetches only what it displays. The tag stream carries both lstar and lstar-sc releases.
The through-line is making a hosted store cheap to read. Sharding addresses the file-per-chunk explosion that makes many-chunk arrays awkward to host; per-field compression with chunk-granular resolution means colouring an embedding by one gene fetches one column rather than an array. The 0.2.x patches are the cost of maintaining four surfaces at once — a WASM heap crash that only browsers exercise, and a count-basis orientation defect where all three surfaces normalized in memory and none owned the on-disk layout.
The orientation bug's root cause — no surface owning the on-disk representation while all three normalized in memory — is the kind of gap that usually produces a validation or ownership change rather than another point fix.
RMediation is a long-standing CRAN package for confidence intervals on mediated effects, now built on an S7 class hierarchy. Over eight weeks it added ProductNormal3 for serial indirect effects of the form a1*a2*b, folded the engine into the existing pprodnormal naming family, and then replaced that engine outright after finding it returned wrong probabilities without warning. The dev branch is at 1.7.0; CRAN still serves 1.6.1.
The package is moving from a hand-rolled numerical layer to one that checks itself: the new default integrator escalates its node count until successive rules agree, warns when it hits the cap instead of returning a number, and exposes a diagnostics argument for the convergence estimate. The correctness fix went to dev ahead of the CRAN window rather than being held for it, which suggests wrong-answer bugs are treated as release-blocking regardless of cadence. Serial mediation is where the new surface area is concentrated.
1.7.0 exists specifically to land before CRAN's 2026-08-21 update window, so the next move is a CRAN submission promoting it to main; whether hcubature survives past that as a cross-check option is the open question.
Other Analytics 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 lstar or rmediation.
The R client for PurpleAir sensors keeps finding its time-averaging was wrong.
A board game graphics package runs one of the most disciplined deprecation cycles in R.
The explainable-ensemble-tree package now measures whether its own explanations are faithful.
The discrete-data FDR package is being pared into one piece of a larger multiple-testing suite.
A scientific-text analysis package moved from counting citations to classifying argument structure.
The teaching arm of an R reliability suite keeps pace with whatever its analysis siblings ship.
See all lstar alternatives → · See all rmediation alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. lstar is currently shipping more aggressively (velocity 5.0 vs 3.8), with 0 editorial sparks in the last 30 days against 1. 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. lstar is currently shipping more aggressively (velocity 5.0 vs 3.8), with 0 editorial sparks in the last 30 days against 1. For your specific use case, the alternatives sections above list other Analytics products to evaluate alongside.
Top lstar alternatives in Analytics are ranked by recent ship velocity. Browse the "lstar alternatives" section above for the current picks, or visit /alternatives/lstar for the full list with editorial commentary on each.
Top rmediation alternatives in Analytics are ranked by recent ship velocity. Browse the "rmediation alternatives" section above for the current picks, or visit /alternatives/rmediation for the full list with editorial commentary on each.