flashlight
flashlight hit 1.0 by giving away its SHAP feature and making most of its API internal.
A side-by-side editorial comparison of fastml and soilDB — release velocity, themes, recent moves, and the top alternatives to consider.
fastml added survival modelling and leakage-proof resampling, moving past classification and regression.
A tidymodels-based AutoML wrapper that trains, tunes and compares many engines from one call. The 0.6.x line added engine-specific tuning parameters, class-imbalance handling, early stopping and DALEX-based explainability. The 0.7.5 release is far larger: a full survival analysis task with its own engines, MICE imputation and integrated Brier scoring, plus unbiased nested cross-validation, grouped, blocked and rolling resampling helpers, fold-wise imputation, recipe leakage checks, and a sandbox for user-supplied preprocessing.
The R front door to USDA soil data finishes a long deprecation cleanup and turns local-first.
soilDB is the R access layer for USDA-NRCS soil data: NASIS local databases, Soil Data Access, SoilWeb coverage services, and a widening set of curated national grids. The 2.9.x line closed out a multi-release deprecation cycle — column aliases and stringsAsFactors are gone, R 4.1 is the floor, and the bundled sample profile collections were rebuilt against the new schema. Recent work has shifted from adding query functions to making existing ones faster and usable against local SQLite or GeoPackage copies.
A tidymodels-based AutoML wrapper that trains, tunes and compares many engines from one call. The 0.6.x line added engine-specific tuning parameters, class-imbalance handling, early stopping and DALEX-based explainability. The 0.7.5 release is far larger: a full survival analysis task with its own engines, MICE imputation and integrated Brier scoring, plus unbiased nested cross-validation, grouped, blocked and rolling resampling helpers, fold-wise imputation, recipe leakage checks, and a sandbox for user-supplied preprocessing.
The package is moving from convenience wrapper to something that has to be defensible statistically. Nested cross-validation, fold-wise rather than up-front imputation, and explicit leakage checks are all corrections to the shortcuts that make AutoML easy and its scores optimistic. Survival adds a third task type alongside classification and regression, and it arrived with its own metrics rather than being bolted onto the existing ones. Note the entry body is cut off at 8,000 characters, so the release is larger than what is shown.
Expect the remaining survival engines to fill in and the sandboxing of custom preprocessing to tighten, since both were still being iterated on within this same release's commit list.
soilDB is the R access layer for USDA-NRCS soil data: NASIS local databases, Soil Data Access, SoilWeb coverage services, and a widening set of curated national grids. The 2.9.x line closed out a multi-release deprecation cycle — column aliases and stringsAsFactors are gone, R 4.1 is the floor, and the bundled sample profile collections were rebuilt against the new schema. Recent work has shifted from adding query functions to making existing ones faster and usable against local SQLite or GeoPackage copies.
The arc points at offline and local-first workflows. downloadSSURGO() and createSSURGO() keep gaining arguments for building and querying local SSURGO databases, and the query internals were rewritten as common table expressions so identical code runs against the remote service or a local file. Coverage is widening in parallel: FY26 SoilWeb maps now reach most OCONUS surveys, while fetchHWSD() and fetchSOLUS() pull in datasets outside the core NASIS/SSURGO pair. Federal URL churn — EDIT, SoilWeb, S3-hosted geometry — is a recurring maintenance tax the package absorbs on users' behalf.
Expect the next releases to keep extending parallel and offline SSURGO handling, since LAPPLY.FUN has just opened the door to arbitrary parallel backends, and to fold more curated SoilWeb and FAO datasets behind fetch* wrappers.
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 fastml or soilDB.
flashlight hit 1.0 by giving away its SHAP feature and making most of its API internal.
A palette generator became a palette platform — and changed the metric behind every color it picks.
An extreme-value toolkit reorganised its whole API into prefixed families and tripled its estimator count.
nflseedR rewrote its simulator from scratch and put the original on a deprecation clock.
The Open Systems Pharmacology plotting layer spent 2025 absorbing a ggplot2 major break.
A frozen astronomical model quietly became the inner loop of its sibling's optimizer.
See all fastml alternatives → · See all soilDB alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. fastml and soilDB 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. fastml and soilDB 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 Analytics products to evaluate alongside.
Top fastml alternatives in Analytics are ranked by recent ship velocity. Browse the "fastml alternatives" section above for the current picks, or visit /alternatives/fastml for the full list with editorial commentary on each.
Top soilDB alternatives in Analytics are ranked by recent ship velocity. Browse the "soilDB alternatives" section above for the current picks, or visit /alternatives/soildb for the full list with editorial commentary on each.