Data Science Tools
Browse all available data science tools.
Available tools
- Batch Drift One Feature Mean Gap Tool Calculate batch drift one feature mean gap from explicit caller-supplied inputs.
- Bootstrap One Draw Multiplicity Term Tool Calculate bootstrap one draw multiplicity term from explicit caller-supplied inputs.
- Categorical Level Reference Label Label an explicit categorical reference level without choosing model coding or interpreting a coefficient.
- Categorical Level Unseen Fallback Token Tool Calculate categorical level unseen fallback token from explicit caller-supplied inputs.
- Cross Validation Purge Window Overlap Tool Half-open label and validation information windows with an explicit embargo appended after validation.
- Data Row Hash Salt Label Label hashing provenance with a nonsecret salt-reference identifier without accepting a salt secret, computing hashes or guaranteeing anonymization.
- Data Split Chronological Cutoff Membership Tool Calculate data split chronological cutoff membership from explicit caller-supplied inputs.
- Dataset Cell Missingness Reason Code Keep an explicitly observed data cell distinct from an allowed missingness reason; observed cells require an empty reason field.
- Dataset Split Single Group Tag Label an explicitly assigned dataset group split without generating a split or validating group isolation.
- Dataset Stratum Sampling Fraction Tool Calculate a supplied dataset-stratum sampling fraction and unsampled count on a positive population-count basis.
- Evaluation Metric Direction Sign Tool Orient a supplied metric change so positive means improvement under an explicit higher-is-better flag.
- Event Time Feature Time Order Flag Compare an explicit feature-availability time with an event boundary under a supplied equality convention without asserting absence of other leakage.
- Feature Availability Lag Gap Tool Calculate feature availability lag gap from explicit caller-supplied inputs.
- Feature Hash Signed Bucket Term Tool Supplied nonnegative hash uses modulo for bucket and parity of the quotient for sign; no hash algorithm is selected.
- Feature Label Time Window Overlap Tool Half-open temporal overlap audit, not a complete information-leakage detector.
- Feature Lag Calendar Row Offset Tool Positional lag on an already ordered calendar-row table; missing calendar periods must be represented explicitly.
- Feature Null Run Length Tool Count one consecutive run in a supplied null mask without treating false, zero or empty string as null.
- Feature Quantization Single Bin Midpoint Tool Return the midpoint of two supplied ordered finite quantization-bin edges without learning bin boundaries.
- Feature Timestamp Latency Budget Gap Tool Calculate supplied feature latency budget minus actual latency without selecting a freshness policy.
- Feature Unit Reference Offset Apply an explicit feature origin and positive input-units-per-output-unit scale without fitting a normalizer.
- Feature Value Unseen Category Flag Identify an exact feature category absent from an explicitly supplied training vocabulary without recoding it.
- Feature Window Availability Slack Tool Calculate feature window availability slack from explicit caller-supplied inputs.
- Fold Group Membership Bit Tool Calculate fold group membership bit from explicit caller-supplied inputs.
- Hash Bucket Sign Collision Term Tool Feature-hashing collision contribution; combined value is unused when supplied buckets differ.
- Label Encoding Cardinality Slack Tool Calculate label encoding cardinality slack from explicit caller-supplied inputs.
- Minmax Feature Train Test Range Gap Tool Apply training anchors without clipping a test observation outside its training range.
- Missingness Indicator One Cell Token Tool Calculate missingness indicator one cell token from explicit caller-supplied inputs.
- Model Residual One Observation Leverage Tool Calculate model residual one observation leverage from explicit caller-supplied inputs.
- Numeric Feature Caller Resolution Bin Tool Calculate numeric feature caller resolution bin from explicit caller-supplied inputs.
- One Hot Reference Level Drop Flag Tool Check whether an explicit one-hot category is the supplied reference level to drop without learning category order.
- Onehot Reference Level Output Budget Tool Calculate onehot reference level output budget from explicit caller-supplied inputs.
- Quantile Bin One Boundary Tie Rule Tool Calculate quantile bin one boundary tie rule from explicit caller-supplied inputs.
- Regression Prediction Retransformation Factor Tool Apply an explicit positive smearing factor after exponentiating a natural-log prediction without estimating residual correction.
- Sample Weight Clipping Mass Loss Tool Calculate sample weight clipping mass loss from explicit caller-supplied inputs.
- Sample Weight Single Effective Mass Return one nonnegative sample weight’s first and second moments after an explicit inclusion mask, without computing a full effective sample size.
- Sample Weight Unknown Support Flag Tool Partition supplied sample weights by independently supplied support-known flags.
- Sparse Feature Implicit Zero Label Tool Record explicit sparse-feature implicit-zero semantics and schema source without treating missing values as zeros.
- Sparse Feature One Entry Density Tool Calculate sparse feature one entry density from explicit caller-supplied inputs.
- Split Entity Support Gap Tool Calculate split entity support gap from explicit caller-supplied inputs.
- Split Group Empty Fold Budget Tool Calculate split group empty fold budget from explicit caller-supplied inputs.
- Standardization Missing Cell Denominator Tool Population variance from explicit nonmissing sufficient statistics; reject materially inconsistent sums rather than treating missing rows as zeros.
- Target Encoding One Out Row Term Tool Leave-one-out target mean with explicit prior pseudo-count; group sum includes the held-out row.
- Target Encoding Prior Single Level Term Tool Calculate an explicit additive-prior target encoding from supplied level sum, count, prior mean and nonnegative pseudo-count.
- Target Label Availability Date Flag Compare an explicit target-label availability date with a supplied training cutoff convention.
- Target Leakage Explicit Time Order Tool One explicit availability-time audit; this does not establish absence of all target leakage.
- Target Observation Window Width Tool Calculate target observation window width from explicit caller-supplied inputs.
- Train Only Scaler Single Value Apply Tool Apply an explicit training-fitted center and positive scale to one supplied value without refitting on evaluation data.
- Train Test Single Identifier Collision Check one exact row identifier for membership in explicit training and test ID sets without detecting broader leakage.
- Train Test Time Gap Tool Compare supplied time-split bounds and an explicit embargo without certifying absence of data leakage.