Skip to contents

Prepare Wealth Quintile Distribution Data for MBG Analysis

Source: R/indicator_wealth.R
calc_wealth_mbg.Rd

Prepares cluster-level wealth quintile distribution data for Model-Based Geostatistics (MBG) analysis. Calculates proportions of households in each wealth quintile, aggregated to cluster level.

Usage

calc_wealth_mbg(
  dhs_hr,
  gps_data,
  indicators = c("prop_poorest", "prop_richest"),
  survey_vars = list(cluster = "hv001", wealth_quintile = "hv270"),
  gps_vars = list(cluster = "DHSCLUST", lat = "LATNUM", lon = "LONGNUM")
)

Arguments

dhs_hr

DHS Household Records (HR) dataset.

gps_data

DHS GPS dataset with cluster coordinates.

indicators

Character vector of indicators to calculate:

  • "prop_poorest" or "prop_q1": Proportion in poorest quintile (Q1)

  • "prop_poorer" or "prop_q2": Proportion in second quintile (Q2)

  • "prop_middle" or "prop_q3": Proportion in middle quintile (Q3)

  • "prop_richer" or "prop_q4": Proportion in fourth quintile (Q4)

  • "prop_richest" or "prop_q5": Proportion in richest quintile (Q5)

Default: c("prop_poorest", "prop_richest") for equity analysis.

survey_vars

Named list mapping DHS variable names:

  • cluster: Cluster ID (default: "hv001")

  • wealth_quintile: Wealth quintile variable (default: "hv270")

gps_vars

Named list for GPS variable mapping.

Value

A named list of data.tables (one per indicator), each with columns:

  • cluster_id: Cluster identifier

  • indicator: Numerator count (households in quintile)

  • samplesize: Denominator count (all households)

  • x: Longitude

  • y: Latitude

Details

This function prepares wealth distribution indicators for spatial modeling. Unlike the survey-weighted calc_wealth_dhs(), this uses simple cluster-level counts without survey weights - MBG handles spatial smoothing internally.

Pipeline Integration: This function IS called by run_mbg_pipeline() when you specify indicators = "wealth" or individual codes like "prop_poorest".

Methodology: Uses DHS wealth quintile variable (hv270 in HR recode) which classifies households into 5 quintiles based on wealth index factor scores.

Output Structure

For indicators = c("prop_poorest", "prop_richest"):


list(
  prop_poorest = data.table(cluster_id, indicator, samplesize, x, y),
  prop_richest = data.table(cluster_id, indicator, samplesize, x, y)
)

See also

Examples

if (FALSE) { # \dontrun{
# Poorest quintile distribution for equity mapping
wealth_poorest <- calc_wealth_mbg(
  dhs_hr = hr_data,
  gps_data = gps_data,
  indicators = "prop_poorest"
)

# Compare poorest vs richest for inequality analysis
wealth_inequality <- calc_wealth_mbg(
  dhs_hr = hr_data,
  gps_data = gps_data,
  indicators = c("prop_poorest", "prop_richest")
)

# Via pipeline
results <- run_mbg_pipeline(
  country_iso3 = "gin",
  indicators = "wealth",
  ...
)
} # }