Skip to contents

Calculate mean (or aggregated) shapes

Source: R/stat_meanshapes.R
stat_meanshape.Rd

Compute mean or other aggregated shapes, either globally or by groups.

Usage

stat_meanshape(data, formula = NULL, .fn = mean, ...)

Arguments

data

A tibble with coefficient columns

formula

A formula or bare column name specifying grouping:

  • Missing/NULL: Global mean across all observations

  • Bare name: status → group by this column, use all coe columns

  • Formula: status ~ . → explicit all coe columns

  • Formula: status ~ VD → specific coe column(s)

  • Formula: ~ status → one-sided formula (group by status)

.fn

Aggregation function. Default mean. Can be median or any function that takes a matrix column and returns a scalar.

...

Additional arguments (reserved for future use)

Value

An object of class c("stat_meanshape", "momstats") containing:

  • data: Original tibble (unchanged)

  • summary: Tibble with group(s), n, and mean coe column(s)

  • method: "meanshape"

  • call: The function call

  • formula: Formula used (if any)

  • group_cols: Grouping column name(s) (NULL if global)

  • coe_cols: Coefficient column names

  • .fn: Aggregation function name

Details

stat_meanshape() provides a unified interface for aggregating shapes. It handles:

  • Global aggregation: No grouping, single mean shape

  • Grouped aggregation: By one or more grouping variables

  • Flexible syntax: Bare names or formula syntax

  • Multiple coe columns: Aggregates all coe columns simultaneously

  • Custom aggregation: mean (default), median, or custom function

  • Automatic inverse: Computes reconstructed shapes via coe_class_i()

Formula syntax

The formula specifies grouping and optionally which coe columns:

  • Missing: Global mean of all coe columns

  • status: Group by status, all coe columns

  • ~ status: Same as above

  • status ~ .: Explicit all coe columns

  • status ~ VD: Only VD coe column

  • status ~ VD + VL: Multiple specific coe columns

Aggregation functions

  • .fn = mean (default): Arithmetic mean

  • .fn = median: Median shape

  • Custom: Any function taking a numeric vector and returning a scalar

Getting results

Use collect() to add mean shapes back to original data (matched by group):

ms <- vitis %>% stat_meanshape(status)
vitis_with_means <- collect(ms)  # Adds mean_VD, mean_VL, etc.

See also

collect.stat_meanshape(), plot.stat_meanshape()

Examples

if (FALSE) { # \dontrun{
# Global mean
ms1 <- vitis %>% stat_meanshape()

# By group
ms2 <- vitis %>% stat_meanshape(status)

# Formula syntax
ms3 <- vitis %>% stat_meanshape(~ status)
ms4 <- vitis %>% stat_meanshape(status ~ .)

# Specific coe column
ms5 <- vitis %>% stat_meanshape(status ~ VD)

# Median instead of mean
ms6 <- vitis %>% stat_meanshape(status, .fn = median)

# Add mean shapes to data
vitis_means <- collect(ms2)

# Plot mean shapes
plot(ms2)  # Mean shapes only
plot(ms2, with_data = TRUE)  # Mean shapes + contributing data
} # }