Wealthier Nations Invest More in Health—But Not Equally

Using wbstats for World Bank data access, faceted scatter plots with population-weighted trend lines, and ggtext for styled annotations.

About the Data

This visualization draws on three key indicators from the World Bank’s World Development Indicators database, accessed through the wbstats R package. The analysis incorporates GDP per capita (current US$) (indicator: NY.GDP.PCAP.CD), which measures a country’s economic output divided by its population; Current health expenditure per capita (current US$) (indicator: SH.XPD.CHEX.PC.CD), reflecting both public and private spending on health services; and Total population (indicator: SP.POP.TOTL), used to weight the analysis and size data points. The dataset spans from 1960 to 2024, though complete data availability varies by country and indicator. Data is aggregated into decade-level averages (2000s, 2010s, 2020s) to smooth annual fluctuations and reveal long-term trends. Continental classifications are derived using the countrycode R package, which maps ISO country codes to geographic regions. Only country-level observations are included, excluding regional aggregates like “World” or “Sub-Saharan Africa” to ensure comparability. The World Bank compiles this data from national statistical offices, international organizations like the World Health Organization, and the International Monetary Fund, making it one of the most comprehensive sources for cross-country development comparisons.

This visualization explores the relationship between national wealth and healthcare investment across 180+ countries over six decades (2000s-2020s). Each dot represents a country’s decade-average values, with dot size proportional to population. The horizontal axis shows GDP per capita on a logarithmic scale, while the vertical axis displays current health expenditure as a percentage of GDP per capita. Colored trend lines reveal continent-specific patterns: as countries grow richer, they generally spend more on healthcare, but this correlation varies significantly. Europe (blue) and the Americas (yellow-green) show stronger positive relationships, while Africa (red) and Asia (teal) display weaker correlations—suggesting that factors beyond economic capacity, such as healthcare policy priorities, governance systems, and cultural values, shape how nations invest in population health.

How I made this graphic?

Loading required libraries, data import & creating custom functions

# Data Import and Wrangling Tools
pacman::p_load(
  tidyverse,    # Data Wrangling and Plotting
  scales,       # Nice scales for ggplot2
  fontawesome,  # Icons display in ggplot2
  ggtext,       # Markdown text support ggplot2
  showtext,     # Display fonts in ggplot2
  colorspace,   # Lighten and darken colours
  patchwork,    # Combining plots together
  magick,       # Image processing and editing
  wbstats,      # World Bank data access
  ggstream,     # Stream Plots in R
  scales        # Nice scales with ggplot2
)

# temp_indicators <- wbstats::wb_indicators()
# 
# temp_indicators |> 
#   as_tibble() |> 
#   filter(str_detect(indicator_desc, "GDP per capita"))
# 
# temp_indicators |> 
#   filter(str_detect(indicator_id, "SH.XPD.TOTL.CD")) |> 
#   select(-indicator_desc, -source_org)


temp_indicators |>
  filter(str_detect(indicator_id, "SH.XPD.CHEX.PC.CD")) |> 
  select(-source_org)

# Source of all indicators: World Development Indicators
selected_indicators <-  c(
      "NY.GDP.PCAP.CD",    # GDP per capita (current US$)
      "SP.POP.TOTL",       # Total population
      "SH.XPD.CHEX.PC.CD"  # Current health expenditure per capita (current US$)
    )


raw_df <- wb_data(
    indicator = selected_indicators,
    start_date = 1960,
    end_date   = 2024,
    # return country-level only (excludes aggregates like "World" if supported)
    country = "countries_only"
  )

Visualization Parameters

# Font for titles
font_add_google("Roboto",
  family = "title_font"
) 

# Font for the caption
font_add_google("Saira Extra Condensed",
  family = "caption_font"
) 

# Font for plot text
font_add_google("Roboto Condensed",
  family = "body_font"
) 

showtext_auto()

# A base Colour
bg_col <- "white"
seecolor::print_color(bg_col)

# Colour for highlighted text
text_hil <- "grey40"
seecolor::print_color(text_hil)

# Colour for the text
text_col <- "grey30"
seecolor::print_color(text_col)

line_col <- "grey30"

# Define Base Text Size
bts <- 80

mypal <- paletteer::paletteer_d("lisa::JackYoungerman") |> 
  as.character() |> 
  str_sub(1,7)

# Caption stuff for the plot
sysfonts::font_add(
  family = "Font Awesome 6 Brands",
  regular = here::here("docs", "Font Awesome 6 Brands-Regular-400.otf")
)
github <- "&#xf09b"
github_username <- "aditya-dahiya"
xtwitter <- "&#xe61b"
xtwitter_username <- "@adityadahiyaias"
social_caption_1 <- glue::glue("<span style='font-family:\"Font Awesome 6 Brands\";'>{github};</span> <span style='color: {text_hil}'>{github_username}  </span>")
social_caption_2 <- glue::glue("<span style='font-family:\"Font Awesome 6 Brands\";'>{xtwitter};</span> <span style='color: {text_hil}'>{xtwitter_username}</span>")
plot_caption <- paste0(
  "**Data:**  World Bank's DataBank",
  " |  **Code:** ",
  social_caption_1,
  " |  **Graphics:** ",
  social_caption_2
)
rm(
  github, github_username, xtwitter,
  xtwitter_username, social_caption_1,
  social_caption_2
)

Annotation Text for the Plot

plot_title <- "The Health-Wealth Paradox: Continental Divides in Healthcare Spending"
str_view(plot_title)

plot_subtitle <- glue::glue(
  "As economies grow, healthcare spending rises as a share of GDP, but the strength of this relationship<br>",
  "varies dramatically by continent. <span style='color:{mypal[4]}'>**European**</span> and ",
  "<span style='color:{mypal[2]}'>**American**</span> nations show the steepest increases, while<br>",
  "<span style='color:{mypal[1]}'>**African**</span> and <span style='color:{mypal[3]}'>**Asian**</span> ",
  "countries lag behind despite economic growth—raising questions about<br>",
  "healthcare priorities, policy choices, and structural inequalities in global health systems."
)

str_view(plot_subtitle)

Exploratory Data Analysis & Data Wrangling

# A tibble of all countries and their GDP per capita and health expenditure per capita
# in current US $ as a percentage of GDP per capita in current US $
df1 <- raw_df |>
  
  # standardize/rename columns and keep necessary cols
  dplyr::rename(
    year = date,
    iso3c = iso3c,
    country = country,
    health_pc = selected_indicators[3],
    gdp_pc = selected_indicators[1],
    pop = selected_indicators[2]
  ) |>
  # add continent using iso3c
  dplyr::mutate(
    continent = countrycode::countrycode(
      iso3c,
      origin = "iso3c",
      destination = "continent",
      warn = FALSE
      ),
    # assign decade label like "1990s", "2000s"
    decade = paste0(floor(as.integer(year) / 10) * 10, "s"),
    health_pc = health_pc / gdp_pc,
  ) |> 
  # keep rows that have all three values - country anme and two indicators
  dplyr::filter(!is.na(gdp_pc) & !is.na(health_pc)) |> 
  select(-iso2c)

# Decade wise summary for countries
df2 <- df1 |> 
  # aggregate per country-decade (mean of available yearly values)
  dplyr::group_by(iso3c, country, continent, decade) |>
  
  dplyr::summarise(
    gdp_pc_mean = mean(gdp_pc, na.rm = TRUE),
    health_pc_mean = mean(health_pc, na.rm = TRUE),
    pop_mean = mean(pop, na.rm = TRUE),
    n_years = sum(!is.na(gdp_pc) | !is.na(health_pc)),
    .groups = "drop"
  ) |>
  # remove groups with missing core values
  dplyr::filter(!is.na(gdp_pc_mean), !is.na(health_pc_mean), n_years > 0) |> 

  group_by(decade) |>
  mutate(
    # Flag top 10 by population in each decade
    top10_pop = rank(-pop_mean) <= 10,
    
    # Calculate distance from cluster center (outliers)
    # Using standardized residuals from the overall trend
    gdp_std = as.vector(scale(log10(gdp_pc_mean))),  # Convert to vector
    health_std = as.vector(scale(health_pc_mean)),    # Convert to vector
    dist_from_center = sqrt(gdp_std^2 + health_std^2),
    is_outlier = dist_from_center > quantile(dist_from_center, 0.85),
    
    # Flag for labeling: top 10 pop OR outliers (roughly 20-25 countries per panel)
    label_this = top10_pop | is_outlier
  ) |>
  ungroup() |> 
  mutate(
    iso2c = countrycode::countrycode(
      iso3c,
      origin = "iso3c",
      destination = "iso2c"
    ),
    iso2c = str_to_lower(iso2c)
  )

The Base Plot

# plotting: log scales on both axes, each decade facet, one point per country
g <- df2 |>
  ggplot(
    mapping = aes(
      x = gdp_pc_mean, 
      y = health_pc_mean
      )
    ) +
  
    geom_point(
      mapping = aes(
        size = sqrt(pop_mean)  # Square root of population
      ),
      alpha = 0.3,
      pch = 19,
      colour = "grey40",  # or your preferred color
      stroke = 0  # Remove borders
    ) +
   
  scale_size(range = c(1, 20)) +
  guides(size = "none") +
  
  geom_smooth(
    mapping = aes(
      group = continent, 
      weight = pop_mean, 
      colour = continent
      ),
    method = "lm",
    span = 1,
    se = FALSE,
    linewidth = 2.5, 
    alpha  = 0.75,
    lineend = "round"
  ) +
  
  
  geom_text(
    data = df2 |> filter(label_this == TRUE),  # Only label selected countries
    mapping = aes(
      label = country
    ),
    nudge_y = 0.003,
    size = bts / 4,
    check_overlap = TRUE,
    colour = text_hil,
    family = "caption_font"
  ) +

    scale_x_log10(
    labels = scales::label_number(
      big.mark = ",",
      scale_cut = cut_short_scale()
    )
  ) +
  scale_y_continuous(
    labels = scales::label_percent()
  ) +
  scale_colour_manual(values = mypal) +
  facet_wrap(~ decade, ncol = 3) +
  
  labs(
    title = plot_title,
    subtitle = plot_subtitle,
    x = "GDP per Capita (current US$) — Decade Mean, Log Scale",
    y = "Health Expenditure as % of GDP per Capita",
    caption = plot_caption
  ) +
  coord_cartesian(
    clip = "off",
    expand = FALSE
  ) +
  theme_minimal(
    base_family = "body_font",
    base_size = bts
  ) +
  theme(
    legend.position = "inside",
    legend.position.inside = c(0,0),
    legend.justification = c(0,1),
    legend.margin = margin(20,0,0,0, "mm"),
    legend.box.margin = margin(0,0,0,0, "mm"),
    legend.direction = "horizontal",
    # legend.key.height = unit(5, "mm"),
    # legend.key.width = unit(15, "mm"),
    legend.text.position = "bottom",
    legend.text = element_text(
      margin = margin(2,0,0,0, "mm"),
      size = bts * 1.1
    ),
    legend.title = element_blank(),
    legend.key.spacing.x = unit(5, "mm"),

    # Overall
    text = element_text(
      margin = margin(0,0,0,0, "mm"),
      colour = text_col,
      lineheight = 0.3
    ),
    
    panel.background = element_rect(
      fill = NA,
      colour = NA
    ),
    panel.grid.major = element_line(
      linewidth = 0.4,
      linetype = 3,
      colour = alpha("black", 0.8)
    ), 
    panel.grid.minor = element_line(
      linewidth = 0.2,
      linetype = 3,
      colour = alpha("black", 0.6)
    ),
    axis.line = element_line(
      linewidth = 0.5,
      colour = text_col,
      arrow = arrow()
    ),
    axis.text.x.bottom = element_text(
      margin = margin(5,1,1,1, "mm" ),
      size = 0.9 * bts
    ),
    axis.text.y.left = element_text(
      margin = margin(2,2,2,2, "mm"),
      size = 1.1 * bts
    ),
    axis.title.x.bottom = element_text(
      margin = margin(2,0,0,0, "mm")
    ),
    axis.title.y.left = element_text(
      margin = margin(0,2,0,0, "mm")
    ),
    strip.text = element_text(
      size = 2 * bts,
      colour = text_hil,
      margin = margin(0,0,0,0, "mm"),
      face = "bold",
      family = "title_font"
    ),
    
    # Labels and Strip Text
    plot.title = element_text(
      margin = margin(5, 0, 5, 0, "mm"),
      hjust = 0.5,
      vjust = 0.5,
      colour = text_hil,
      size = 2 * bts,
      family = "title_font",
      face = "bold",
      lineheight = 0.25
    ),
    plot.subtitle = element_textbox(
      margin = margin(2, 0, 2, 0, "mm"),
      vjust = 0.5,
      colour = text_hil,
      size = 1.15 * bts,
      hjust = 0,
      halign = 0,
      family = "body_font",
      lineheight = 0.3
    ),
    plot.caption = element_markdown(
      family = "caption_font",
      hjust = 1,
      margin = margin(5,0,0,0, "mm"),
      colour = text_hil
    ),
    plot.caption.position = "plot",
    plot.title.position = "plot",
    plot.margin = margin(5, 5, 5, 5, "mm")
  )

Adding annotations to the plot

# A QR Code for the infographic
url_graphics <- paste0(
  "https://aditya-dahiya.github.io/projects_presentations/data_vizs/",
  # The file name of the current .qmd file
  "wb_health_exp_gdp",         
  ".html"
)
# remotes::install_github('coolbutuseless/ggqr')
# library(ggqr)
plot_qr <- ggplot(
  data = NULL, 
  aes(x = 0, y = 0, label = url_graphics)
  ) + 
  ggqr::geom_qr(
    colour = text_hil, 
    fill = bg_col,
    size = 1.5
    ) +
  annotate(
    geom = "text",
    x = -0.08,
    y = 0,
    label = "Scan for complete\nCode used to make\nthis graphic",
    hjust = 1,
    vjust = 0.5,
    family = "caption_font",
    colour = text_hil,
    size = bts / 6,
    lineheight = 0.35
  ) +
  coord_fixed(clip = "off") +
  theme_void() +
  theme(
    plot.background = element_rect(
      fill = NA, 
      colour = NA
    ),
    panel.background = element_rect(
      fill = NA,
      colour = NA
    ),
    plot.margin = margin(0, 10, 0, 0, "mm")
  )

# Compiling the plots

g_full <- g +
  inset_element(
    p = plot_qr,
    left = 0.92, right = 0.98,
    bottom = 0.84, top = 0.9,
    align_to = "full",
    clip = FALSE
  ) + 
  plot_annotation(
    theme = theme(
      plot.background = element_rect(
        fill = "transparent",
        colour = "transparent"
      )
    )
  )

ggsave(
  filename = here::here(
    "data_vizs",
    "wb_health_exp_gdp.png"
  ),
  plot = g_full,
  width = 297 * 2,
  height = 210 * 2,
  units = "mm",
  bg = bg_col
)

Savings the graphics

# Saving a thumbnail for the webpage
image_read(here::here("data_vizs", "wb_health_exp_gdp.png")) |> 
  image_resize(geometry = "400") |> 
  image_write(here::here("data_vizs", 
                         "thumbnails", 
                         "wb_health_exp_gdp.png"))

Session Info

# Data Import and Wrangling Tools
pacman::p_load(
  tidyverse,    # Data Wrangling and Plotting
  scales,       # Nice scales for ggplot2
  fontawesome,  # Icons display in ggplot2
  ggtext,       # Markdown text support ggplot2
  showtext,     # Display fonts in ggplot2
  colorspace,   # Lighten and darken colours
  patchwork,    # Combining plots together
  magick,       # Image processing and editing
  wbstats,      # World Bank data access
  ggstream,     # Stream Plots in R
  scales        # Nice scales with ggplot2
)

sessioninfo::session_info()$packages |>
  as_tibble() |>
  
  # The attached column is TRUE for packages that were 
  # explicitly loaded with library()
  dplyr::filter(attached == TRUE) |>
  dplyr::select(package,
    version = loadedversion,
    date, source
  ) |>
  dplyr::arrange(package) |>
  janitor::clean_names(
    case = "title"
  ) |>
  gt::gt() |>
  gt::opt_interactive(
    use_search = TRUE
  ) |>
  gtExtras::gt_theme_espn()

Table 1: R Packages and their versions used in the creation of this page and graphics