Automatically selects and generates the most appropriate statistical visualization based on variable data types. Features enhanced error messages with contextual guidance, robust data validation, and comprehensive fallback options. Supports both independent and repeated measures designs with various plot types including violin plots, scatter plots, bar charts, and alluvial diagrams.
Usage
statsplot2(
data,
dep = NULL,
group = NULL,
grvar = NULL,
direction = "independent",
distribution = "p",
alluvsty = "t1",
excl = FALSE,
sampleLarge = FALSE,
sampleThreshold = 10000,
sampleSize = 5000,
seed = 42
)Arguments
- data
The data as a data frame.
- dep
The dependent variable (y-axis, 1st measurement). Can be continuous or categorical.
- group
The grouping variable (x-axis, 2nd measurement). Can be continuous or categorical.
- grvar
Optional grouping variable for creating grouped plots across multiple panels.
- direction
Measurement design type. "independent" for between-subjects comparisons, "repeated" for within-subjects/repeated measures comparisons.
- distribution
Statistical approach: "p" = parametric, "np" = nonparametric, "r" = robust, "bf" = Bayes factor.
- alluvsty
Style for alluvial diagrams: "t1" = ggalluvial with stratum labels, "t2" = easyalluvial with automatic variable selection.
- excl
If TRUE, excludes rows with missing values before analysis.
- sampleLarge
If TRUE, automatically samples large datasets (>10,000 rows) to 5,000 rows for improved performance.
- sampleThreshold
Row count above which sampling is applied, when 'Sample large datasets' is enabled. Datasets at or below this size are always analysed in full.
- sampleSize
Number of rows to retain when sampling. Larger values keep more statistical power at the cost of plotting speed. Statistics are computed on the retained rows only.
- seed
Random seed for the reproducible sampling of large datasets (used when 'Sample Large Datasets' is enabled). Change it to draw a different sample; the default (42) reproduces the previous fixed behaviour.