mbg (version 1.1.0)

summarize_draws: Summarize draws

Description

Helper function to summarize a matrix or data.frame of predictive draws

Usage

summarize_draws(
  draws,
  id_fields = NULL,
  draw_fields = NULL,
  ui_width = 0.95,
  na.rm = TRUE
)

Value

A data.table::data.table containing at least the following fields:

• The id_fields, if passed

• "mean": Mean across predictive draws

• "lower": Lower bound of the (X%) uncertainty interval

• "upper": Upper bound of the (X%) uncertainty interval

• "ui_width": "upper" - "lower"

Arguments

draws

A matrix, data.frame, or data.table::data.table of predictive draws.

id_fields

(default NULL) Only considered for data.frame-like draws. What identifier fields in the data should be kept in the summary table and not included among the draw fields?

draw_fields

(default NULL) Only considered for data.frame-like draws. What fields represent actual draws, as opposed to identifier fields or other metadata like population? If NULL, the default, automatically determines the draw fields as all columns not included in the id_fields.

ui_width

(numeric, default 0.95) Size of the uncertainty interval width when calculating the upper and lower summary rasters

na.rm

(logical, default TRUE) Should NA values be removed when calculating summaries across draws?

Examples

# Summarize a draws matrix
draws_matrix <- matrix(rnorm(200), nrow = 10)
summary_table_a <- summarize_draws(draws_matrix)
head(summary_table_a)

# Summarize a draws data.table with location IDs
draws_table <- matrix(c(1:10, rnorm(200)), nrow = 10) |>
  data.table::as.data.table() |>
  data.table::setnames(c('location_id', paste0('draw_', 1:20)))
summary_table_b <- summarize_draws(draws_table, id_fields = 'location_id')
head(summary_table_b)