Skip to contents

Summarize Measurements Across Samples

Source: R/measure-map.R
measure_summarize.Rd

measure_summarize() computes summary statistics for each measurement location across all samples. This is useful for understanding your data, computing reference spectra, or identifying outliers.

Usage

measure_summarize(
  .data,
  .cols = NULL,
  .fns = list(mean = mean, sd = stats::sd),
  na.rm = TRUE
)

Arguments

.data

A data frame containing one or more measure_list columns.

.cols

<tidy-select> Columns to summarize. Defaults to all measure_list columns.

.fns

A named list of summary functions. Each function should accept a numeric vector and return a single value. Default is list(mean = mean, sd = sd).

na.rm

Logical. Should NA values be removed? Default is TRUE.

Value

A tibble with one row per measurement location and columns for each summary statistic.

Details

This function does NOT transform data; it summarizes it. Common uses:

  • Mean spectrum: The average spectrum across all samples

  • Reference spectrum: For MSC-style corrections

  • Variability: Standard deviation at each wavelength

  • Quality control: Identify problematic wavelength regions

Examples

library(recipes)
library(ggplot2)

rec <- recipe(water + fat + protein ~ ., data = meats_long) |>
  update_role(id, new_role = "id") |>
  step_measure_input_long(transmittance, location = vars(channel)) |>
  prep()

baked_data <- bake(rec, new_data = NULL)

# Compute mean and SD at each wavelength
summary_stats <- measure_summarize(baked_data)
summary_stats
#> # A tibble: 100 × 3
#>    location  mean    sd
#>       <int> <dbl> <dbl>
#>  1        1  2.81 0.411
#>  2        2  2.81 0.413
#>  3        3  2.81 0.416
#>  4        4  2.82 0.418
#>  5        5  2.82 0.421
#>  6        6  2.82 0.424
#>  7        7  2.83 0.426
#>  8        8  2.83 0.429
#>  9        9  2.83 0.432
#> 10       10  2.84 0.434
#> # ℹ 90 more rows

# Visualize mean spectrum with confidence band
ggplot(summary_stats, aes(x = location)) +
  geom_ribbon(aes(ymin = mean - sd, ymax = mean + sd), alpha = 0.3) +
  geom_line(aes(y = mean)) +
  labs(x = "Channel", y = "Transmittance", title = "Mean Spectrum +/- 1 SD")


# Custom summary functions
measure_summarize(
  baked_data,
  .fns = list(
    median = median,
    q25 = function(x) quantile(x, 0.25),
    q75 = function(x) quantile(x, 0.75)
  )
)
#> # A tibble: 100 × 4
#>    location median   q25   q75
#>       <int>  <dbl> <dbl> <dbl>
#>  1        1   2.75  2.51  3.01
#>  2        2   2.76  2.51  3.01
#>  3        3   2.76  2.51  3.01
#>  4        4   2.76  2.52  3.02
#>  5        5   2.76  2.52  3.03
#>  6        6   2.76  2.52  3.03
#>  7        7   2.76  2.52  3.04
#>  8        8   2.77  2.52  3.05
#>  9        9   2.77  2.52  3.05
#> 10       10   2.77  2.52  3.06
#> # ℹ 90 more rows