Which municipalities improved the most in IDEB?

This vignette shows how to use educabR to identify which municipalities had the largest IDEB improvement over time, using the historical series available via get_ideb().

library(educabR)
library(dplyr)
library(tidyr)
library(ggplot2)

Downloading IDEB data

get_ideb() returns data in long format with all available editions. A single call with year = c(2005, 2023) gives us the two endpoints we need for comparison.

ideb <- get_ideb(
  level  = "municipio",
  stage  = "anos_iniciais",
  metric = "indicador",
  year   = c(2005, 2023)
)
glimpse(ideb)

#> Rows: 86,982
#> Columns: 7
#> $ uf_sigla         <chr> "RO", "RO", "RO", "RO", "RO", "RO", …
#> $ municipio_codigo <chr> "1100015", "1100015", "1100015", "1100…
#> $ municipio_nome   <chr> "Alta Floresta D'Oeste", "Alta Florest…
#> $ rede             <chr> "Estadual", "Municipal", "Pública", "E…
#> $ ano              <int> 2005, 2005, 2005, 2005, 2005, 2005, 20…
#> $ indicador        <chr> "IDEB", "IDEB", "IDEB", "IDEB", "IDEB…
#> $ valor            <dbl> 3.5, NA, 3.7, 4.0, 3.5, 3.7, 4.1, 3.4…

Calculating IDEB improvement

We filter to the IDEB indicator for municipal schools, pivot to wide format to get one column per year, then compute the change.

ideb_change <-
  ideb |>
  filter(indicador == "IDEB") |>
  filter(rede == "Municipal") |>
  select(uf_sigla, municipio_codigo, municipio_nome, ano, valor) |>
  pivot_wider(names_from = ano, values_from = valor, names_prefix = "ideb_") |>
  filter(!is.na(ideb_2005), !is.na(ideb_2023)) |>
  mutate(
    change     = ideb_2023 - ideb_2005,
    pct_change = change / ideb_2005 * 100
  )

Top 20 municipalities by absolute improvement

ideb_change |>
  slice_max(change, n = 20) |>
  ggplot(aes(
    x = reorder(paste(municipio_nome, uf_sigla, sep = " - "), change),
    y = change
  )) +
  geom_col(fill = "#2a9d8f") +
  coord_flip() +
  labs(
    title    = "Top 20 Municipalities by IDEB Improvement (2005-2023)",
    subtitle = "Municipal schools, early elementary (1st-5th grade)",
    x        = NULL,
    y        = "IDEB change (2023 - 2005)"
  ) +
  theme_minimal()

Distribution of improvement by state

ideb_change |>
  ggplot(aes(x = reorder(uf_sigla, change, FUN = median), y = change)) +
  geom_boxplot(fill = "#2a9d8f", alpha = 0.6, outlier.size = 0.5) +
  coord_flip() +
  labs(
    title    = "IDEB Improvement Distribution by State (2005-2023)",
    subtitle = "Municipal schools, early elementary (1st-5th grade)",
    x        = NULL,
    y        = "IDEB change"
  ) +
  theme_minimal()

IDEB trajectory for selected municipalities

# Pick the top 5 improvers
top5 <-
  ideb_change |>
  filter(rede == "Municipal") |>
  slice_max(change, n = 5) |>
  pull(municipio_codigo)

# Get full history for these municipalities
ideb_full <-
  get_ideb(
    level  = "municipio",
    stage  = "anos_iniciais",
    metric = "indicador"
  ) |>
  filter(rede == "Municipal")

ideb_full |>
  filter(municipio_codigo %in% top5, indicador == "IDEB") |>
  mutate(label = paste(municipio_nome, uf_sigla, sep = " - ")) |>
  ggplot(aes(x = factor(ano), y = valor, color = label, group = label)) +
  geom_line(linewidth = 1) +
  geom_point(size = 2) +
  labs(
    title    = "IDEB Trajectory - Top 5 Improvers",
    subtitle = "Municipal schools, early elementary (1st-5th grade)",
    x        = "Year",
    y     = "IDEB",
    color = NULL
  ) +
  theme_minimal() +
  theme(legend.position = "bottom")

Municipalities that declined

ideb_change |>
  filter(change < 0) |>
  slice_min(change, n = 15) |>
  ggplot(aes(
    x = reorder(paste(municipio_nome, uf_sigla, sep = " - "), change),
    y = change
  )) +
  geom_col(fill = "#e76f51") +
  coord_flip() +
  labs(
    title    = "Municipalities with Largest IDEB Decline (2005-2023)",
    subtitle = "Municipal schools, early elementary (1st-5th grade)",
    x        = NULL,
    y        = "IDEB change (2023 - 2005)"
  ) +
  theme_minimal()