ggaction 0.0.12

Statistical Data Transforms

Cars scatterplot with grouped regression lines and confidence bands
Layered points, derived fits, and confidence bands.

createRegressionData({ id, source?, x, y, groupBy?, method?, degree?, span?, confidence?, interval? })

Create deterministic regression predictions from an existing dataset. This is an advanced data action used by higher-level regression chart actions.

program.createRegressionData({
  id: "regressionData",
  x: "Displacement",
  y: "Acceleration",
  groupBy: "Origin"
});
Option Type Default
id new dataset ID required
source existing dataset ID current dataset
x, y finite quantitative field names required
groupBy nominal field name one ungrouped model
method "linear", "polynomial", or "loess" "linear"
degree integer from 1 through 32 for polynomial 2
span number greater than 0 and at most 1 for LOESS 0.75
confidence number strictly between 0 and 1; not accepted by LOESS 0.95
interval "mean" or "prediction"; not accepted by LOESS "mean"

Linear and polynomial fits use stable least squares. Polynomial groups require at least degree + 2 rows and degree + 1 distinct x values. LOESS uses tricube-weighted local-linear fits over ceil(span * groupSize) nearest rows, with source order breaking distance ties. Output follows group first appearance and observed unique x order. Linear and polynomial output includes fixed __regression_ci_lower / __regression_ci_upper fields; LOESS is line-only. Source values remain unchanged.

Regression output is limited to 10,000 unique group/x rows. Polynomial and LOESS fitting reject inputs whose estimated work exceeds 10,000,000 units, before allocating or entering the expensive fit.

createIntervalData({ id, source?, field, groupBy?, center?, extent?, level?, as? })

Create immutable grouped interval-summary rows independently from an error-bar mark.

program.createIntervalData({
  id: "accelerationIntervals",
  field: "Acceleration",
  groupBy: "Origin"
});
Option Type Default
id new dataset ID required
source existing dataset ID current dataset
field quantitative field name required
groupBy field name or array of field names one ungrouped interval
center "mean" or "median" "mean"
extent "stderr", "stdev", "ci", or "iqr" "ci"
level number strictly between 0 and 1 0.95 for CI
as { center, lower, upper } distinct field names ID-namespaced fields

Mean supports standard error, sample standard deviation, and two-sided Student-t confidence intervals. Median requires interquartile range, and IQR requires median. Group order follows first appearance. Missing group values, non-finite measures, and undersized mean groups are omitted; valid source rows and the source dataset remain unchanged. A summary whose derived finite center or endpoint is not representable throws a RangeError atomically.

createDensityData({ id, source?, field, groupBy?, bandwidth?, extent?, steps?, kernel?, normalization?, as? })

Create an immutable kernel-density dataset. This is an advanced data action used by higher-level density-chart encodings.

program.createDensityData({
  id: "accelerationDensity",
  field: "Acceleration",
  groupBy: "Origin",
  bandwidth: 0.6
});
Option Type Default
id new dataset ID required
source existing dataset ID current dataset
field quantitative field name required
groupBy nominal field name one ungrouped density
bandwidth positive number or "auto" automatic Scott-rule estimate
extent ascending finite pair or "auto" observed valid extent
steps integer from 2 through 10,000 100
kernel "gaussian", "epanechnikov", "uniform", or "triangular" "gaussian"
normalization "unit" or "count" "unit"
as two distinct output field names <field>_value, <field>_density

Grouped densities use one shared extent and inclusive sample grid. Group order follows first appearance. Transform provenance keeps the requested bandwidth and extent values—including "auto"—and stores the materialized revision’s concrete values separately as resolved: { bandwidth, extent }. Kernel and normalization defaults are stored directly. Unit normalization integrates each complete group density to one; count normalization scales it by that group’s valid sample count. Density output is limited to 10,000 rows, and valid source rows multiplied by steps must not exceed the 10,000,000-unit work budget. Source values remain unchanged.

Data overview · Chart API · Action reference