Working with PNAD Housing

pnad_housing provides annual household estimates from five PNAD Contínua SIDRA tables. The default table describes housing tenure.

tenure <- get_dataset("pnad_housing")

rented_share <- tenure |>
  dplyr::filter(
    geography_type == "brazil",
    category == "rented",
    variable == "household_share"
  )

Counts remain in thousand households. Shares and coefficients of variation remain percentages on a 0-100 scale. Each table includes its total category; exclude category == "total" before summing component categories.

dwelling_type <- get_dataset("pnad_housing", table = "dwelling_type")

apartment_share <- dwelling_type |>
  dplyr::filter(
    geography_type == "brazil",
    category == "apartment",
    variable == "household_share"
  )

The housing-module tables publish 2016-2019 and 2022 onward. They contain no 2020 or 2021 observations. Domestic-unit composition has a longer annual series and describes relationships to the reference person; it is not a count of families.

household_size <- get_dataset("pnad_housing", table = "household_size")

one_person <- household_size |>
  dplyr::filter(
    geography_type == "brazil",
    category == "one_resident",
    variable %in% c("households", "household_share")
  )

mean_size <- get_dataset("pnad_housing", table = "mean_household_size") |>
  dplyr::filter(
    geography_type == "brazil",
    variable == "mean_residents"
  )

The composition table publishes counts and their CVs. A composition share can be calculated from a category count and the same-table total, but the count CV is not the CV of that derived share.