Simulate Ready-to-Use Statistical Datasets

Generate synthetic datasets suitable for common statistical analyses. Each dataset type contains real signal so the intended analysis works out of the box. Different seeds produce structurally different datasets (varying n, effect sizes, number of groups/variables).

Usage

simulate_data(type, seed = NULL, complexity = "clean", ..., n_batch = NULL)

Arguments

type

Character string specifying the dataset type. One of:

"ttest"

Two-group comparison. Columns: group (factor), score (numeric).

"anova"

Multi-group comparison (3–5 groups). Columns: group (factor), score (numeric).

"correlation"

Correlated variables (4–7 columns). All numeric x1–xp.

"clusters"

Cluster structure. Numeric x1–xd plus integer true_cluster.

"factor_analysis"

Latent factor structure. All numeric x1–xp. Attributes n_factors and loadings.

"prediction"

Regression dataset. Columns: y, x1–x4, cat1, cat2.

"mlvar"

Multilevel VAR panel data. Columns: id, day, beep, V1–Vd. Attributes true_temporal, true_contemporaneous, vars.

"batch"

Wildcard: generates all 7 types. Returns a named list with n_batch (default 1000) datasets per type.

seed

Integer or NULL. Random seed. Also determines structural parameters (n, effect sizes, etc.). Default: NULL.

complexity

Character. Controls edge-case injection for stress testing.

"clean"

(default for single datasets) No edge cases.

"auto"

(default in batch mode) Randomly injects 0–3 edge cases per dataset (seed-driven, fully reproducible).

character vector

Inject specific cases. One or more of: "na", "outliers", "ties", "duplicates", "constant_col", "all_na_col", "tiny_n", "heavy_tailed", "heteroscedastic", "extreme_imbalance", "multicollinear".

...

Optional overrides for structural parameters (n, n_groups, etc.).

n_batch

Integer or NULL. When provided, returns a list of n_batch datasets instead of a single dataset. When type = "batch", defaults to 1000L. Each item has seed, batch_id, and complexity attributes.

Value

A data.frame (single dataset) or a list (when n_batch is non-NULL or type = "batch"). Each dataset has attributes type, info, and complexity. Batch items additionally have seed and batch_id.

See also

simulate_sequences(), simulate_igraph()

Examples

# Single clean dataset
d <- simulate_data("ttest", seed = 42)

# With edge cases
d <- simulate_data("correlation", seed = 1, complexity = "auto")
d <- simulate_data("ttest", seed = 1, complexity = c("na", "outliers"))

# Batch of 100 ttest datasets
batch <- simulate_data("ttest", seed = 1, n_batch = 100)

# All types, 1000 datasets each
all_batches <- simulate_data("batch", seed = 1)