The geobr package provides quick and easy access to official spatial data sets of Brazil. The package offers a wide range of spatial data sets available at various geographic scales and for various years with harmonized attributes, projection and fixed topology. All geobr functions follow a simple and consistent syntax that allows users to seamlessly download data and work with it either in memory using sf or out of memory using DuckDB and Arrow. This vignette presents a quick intro to geobr.

Installation

You can install geobr from CRAN or the development version to use the latest features.

# From CRAN
install.packages("geobr")

# Development version
utils::remove.packages('geobr')
devtools::install_github("ipeaGIT/geobr", subdir = "r-package")

Now let’s load the libraries we’ll use in this vignette.

library(geobr)
library(sf)
library(dplyr)
library(ggplot2)

General usage

Available data sets

The geobr package currently covers 30 spatial data sets, including a variety of political-administrative and statistical areas used in Brazil. You can view what data sets are available using the list_geobr() function.

# Available data sets
datasets <- list_geobr(wide = TRUE)

head(datasets)

Basic syntax

The syntax of all geobr functions operate on the same simple logic, so the code to download the data becomes intuitive for the user. Here are a few examples.

Download an specific geographic area at a given year:

# State of Sergipe
state <- read_state(
  year = 2022,
  code_state = "SE",
  showProgress = FALSE
  )

# Municipality of Sao Paulo
muni <- read_municipality(
  year = 2022, 
  code_muni = 3550308, 
  showProgress = FALSE
  )

ggplot() + 
  geom_sf(data = muni, color=NA, fill = '#1ba185') +
  theme_void()

Download all geographic areas within a state at a given year:

# All municipalities in the state of Minas Gerais
muni <- read_municipality(
  year = 2022,
  code_muni = "MG", 
  showProgress = FALSE
  )

head(muni)

If the parameter code_ is not passed to the function, geobr returns the data for the whole country by default.

# read all schools
inter <- read_schools(
  year = 2022,
  showProgress = FALSE
  )

# read all states
states <- read_state(
  year = 2025, 
  showProgress = FALSE
  )

head(states)

Important note about data resolution

All functions to download polygon data such as states, municipalities etc. have a simplified argument. When simplified = FALSE, geobr returns the original data set with high resolution at detailed geographic scale (see documentation). By default, however, simplified = TRUE and geobr returns data geometries with simplified borders to improve speed of downloading and plotting the data.

Plot the data

Once you’ve downloaded the data, it is really simple to plot maps using ggplot2.

# Remove plot axis
no_axis <- theme(axis.title=element_blank(),
                 axis.text=element_blank(),
                 axis.ticks=element_blank())

# Plot all Brazilian states
ggplot() +
  geom_sf(data=states, fill="#2D3E50", color="#FEBF57", size=.15, show.legend = FALSE) +
  labs(subtitle="States", size=8) +
  theme_minimal() +
  no_axis

Plot all the municipalities of a particular state, such as Rio de Janeiro:


# Download all municipalities of Rio
all_muni <- read_municipality(
  year= 2022,
  code_muni = "RJ", 
  showProgress = FALSE
  )

# plot
ggplot() +
  geom_sf(data=all_muni, fill="#2D3E50", color="#FEBF57", size=.15, show.legend = FALSE) +
  labs(subtitle="Municipalities of Rio de Janeiro, 2000", size=8) +
  theme_minimal() +
  no_axis

Lazy evaluation with DuckDB and Arrow

By default, all functions in geobr use output = "sf" and return sf objects loaded into memory. In some cases, however, it may be preferable to process data out of memory for faster and more memory-efficient computation, particularly when working with large spatial data sets.

To support these workflows, users can set output = "duckdb" to return a lazy duckspatial_df object. This allows data to be analyzed with DuckDB using the {duckspatial} package, enabling efficient out-of-memory spatial operations using a syntax similar to {sf}.

Alternatively, users can set output = "arrow" to return an Arrow dataset, which can be integrated with the Arrow ecosystem for scalable analytical workflows.

# return duckdb duckspatial_df
muni_duck <- geobr::read_municipality(
  year = 2022, 
  output = "duckdb"
  )

# return arrow table
muni_arrow <- geobr::read_municipality(
  year = 2022, 
  output = "arrow"
  )

Thematic maps

The next step is to combine data from geobr package with other data sets to create thematic maps. In this first example, we will be using data from the (Atlas of Human Development (by Ipea/FJP and UNPD) to create a choropleth map showing the spatial variation of Life Expectancy at birth across Brazilian states.

Merge external data

First, we need a data.frame with estimates of Life Expectancy. We then need to merge this table to our spatial database. The two-digit abbreviation of state name is our key column to join these two data sets.

# Read data.frame with life expectancy data
df <- data.table::fread(
  system.file("extdata/br_states_lifexpect2017.csv", package = "geobr")
  )

# join the databases
states <- dplyr::left_join(
  x = states, 
  y = df, 
  by = c("name_state" = "uf")
  )

Plot thematic map

ggplot() +
  geom_sf(data=states, aes(fill=ESPVIDA2017), color= NA, size=.15) +
    labs(subtitle="Life Expectancy at birth, Brazilian States, 2014", size=8) +
    scale_fill_distiller(palette = "Blues", name="Life Expectancy", limits = c(65,80)) +
    theme_minimal() +
    no_axis

Using geobr together with censobr

Following the same steps as above, we can use together geobr with our sister package censobr to map the proportion of households connected to a sewage network in Brazilian municipalities

First, we need to download households data from the Brazilian census using the read_households() function.

library(censobr)
library(arrow)

hs <- read_households(
  year = 2010, 
  showProgress = FALSE
  )

Now we’re going to (a) group observations by municipality, (b) get the number of households connected to a sewage network, (c) calculate the proportion of households connected, and (d) collect the results.

esg <- hs |> 
        collect() |>
        group_by(code_muni) |>                                             # (a)
        summarize(rede = sum(V0010[which(V0207=='1')]),                    # (b)
                  total = sum(V0010)) |>                                   # (b)
        mutate(cobertura = rede / total) |>                                # (c)
        collect()                                                          # (d)

head(esg)

Now we only need to download the geometries of Brazilian municipalities from geobr, merge the spatial data with our estimates and map the results.

# download municipality geometries
muni_sf <- geobr::read_municipality(
  year = 2010,
  showProgress = FALSE
  )

# merge data
esg_sf <- left_join(muni_sf, esg, by = 'code_muni')

# plot map
ggplot() +
  geom_sf(data = esg_sf, aes(fill = cobertura), color=NA) +
  labs(title = "Share of households connected to a sewage network") +
  scale_fill_distiller(palette = "Greens", direction = 1, 
                       name='Share of\nhouseholds', 
                       labels = scales::percent) +
  theme_void()

Introductio to geobr (R)