Skip to contents

Process Weighted Raster Data in Batch

Source: R/process_rasters.R
process_weighted_raster_collection.Rd

Processes multiple raster files with population weights, matching value rasters with corresponding population rasters by year.

Usage

process_weighted_raster_collection(
  value_raster_dir,
  pop_raster_dir,
  shapefile,
  id_cols = c("adm0", "adm1", "adm2"),
  value_pattern = "\\.tiff$",
  pop_pattern = "\\.tif$",
  value_layer_to_process = 1,
  weight_na_as_zero = TRUE,
  stat_type = "mean",
  year_extractor = NULL
)

Arguments

value_raster_dir

Character. Directory containing value raster files

pop_raster_dir

Character. Directory containing population raster files

shapefile

sf/SpatVector object. Shapefile for spatial aggregation

id_cols

Character vector. Column names for spatial identifiers (default: c("adm0", "adm1", "adm2"))

value_pattern

Character. Regex pattern to match value raster files (default: "\.tiff$")

pop_pattern

Character. Regex pattern to match population raster files (default: "\.tif$")

value_layer_to_process

Integer or character. Specifies which layer in a multi-layer raster to extract. If the raster contains multiple layers (e.g., different years or indicators), this argument selects the layer to be processed. Default is 1.

weight_na_as_zero

Logical. Treat NA weights as 0 (default: TRUE)

stat_type

Character string specifying statistic type:

  • "mean": Population-weighted mean using exactextractr's weighted_mean function

  • "median": Population-weighted median using matrixStats::weightedMedian() approach that finds the value where 50% of the population weight lies below and 50% above, interpolating between pixel values when needed

  • "both": Returns both weighted mean and weighted median statistics (default: "mean")

year_extractor

Optional function. A custom function that accepts a filename (character) and returns a year as integer or character. Passed through to extract_time_components(). Use this when the default "last plausible 4-digit year in 1980-2099" heuristic picks the wrong year (for example with malariaAtlas filenames that contain multiple year-like tokens). Defaults to NULL, which uses the built-in extractor.

Value

A data frame containing:

  • Spatial identifiers from id_cols

  • Weighted mean values

  • Metadata (filename, year, month, semester, quarter, date)

Examples

if (FALSE) { # \dontrun{
results <- process_weighted_raster_collection(
  value_raster_dir = "path/to/values",
  pop_raster_dir = "path/to/population",
  shapefile = my_shapefile,
  id_cols = c("adm0", "adm1")
)
} # }