Country metadata you don’t have to maintain

A surprising amount of global-health code starts by hand-maintaining a list of countries, regions, and ISO codes. DSIR ships that as data. who_countries is a tibble of all 194 WHO Member States, with ISO3 / ISO2 / UN M49 codes, official and short names, WHO region, and a flag for Pacific Island Countries:

library(DSIR)
library(dplyr)

who_countries |>
  filter(who_region == "WPR") |>
  select(iso3, name_short, is_pic)

# A tibble: 28 × 3
   iso3  name_short        is_pic
   <chr> <chr>             <lgl> 
 1 AUS   Australia         FALSE 
 2 BRN   Brunei Darussalam FALSE 
 3 KHM   Cambodia          FALSE 
 4 CHN   China             FALSE 
 5 COK   Cook Islands      TRUE  
 6 FJI   Fiji              TRUE  
 7 IDN   Indonesia         FALSE 
 8 JPN   Japan             FALSE 
 9 KIR   Kiribati          TRUE  
10 LAO   Lao PDR           FALSE 
# ℹ 18 more rows

There are also ready-made ISO3 vectors for each region — wpro_cty, afro_cty, euro_cty, and so on — that you can pass straight into the data functions. No more pasting country lists between scripts.

A consistent “search → fetch → clean” rhythm

Both API clients follow the same three-step rhythm, so once you’ve learned one you’ve learned both.

For GHO, you can search for an indicator, check coverage before committing to a download, then fetch and tidy:

# Search indicators by keyword
gho_indicators("mortality")

# Fetch premature NCD mortality for the Western Pacific, then tidy
raw <- gho_data("NCDMORT3070", spatial_type = "country", area = wpro_cty)
gho_clean(raw)

The SDG client works exactly the same way — and crucially, its area argument also accepts ISO3 codes, converting to M49 internally, so the same regional vectors just work:

# 3.4.1 = NCD mortality; pass ISO3 codes directly
raw <- sdg_data("3.4.1", area = wpro_cty)
sdg_clean(raw)

The payoff: one schema, so you can just stack them

Here is the part that saves the most time. gho_clean() and sdg_clean() both return the same 15-column schema (source, id, indicator, location, iso3, year, value, value_num, …). Because the shape is identical, GHO and SDG output combine directly — no manual renaming, no reconciling code systems:

gho <- gho_data("NCDMORT3070", area = wpro_cty) |> gho_clean()
sdg <- sdg_data("3.4.1",        area = wpro_cty) |> sdg_clean()

bind_indicators(gho, sdg)   # keep track of origin via the `source` column

That source column means you never lose track of where a row came from, even after you’ve stacked half a dozen indicators from both APIs.

Charts and tables that already look like WHO

The other recurring cost in this kind of work is making every chart and table look consistent across a report. DSIR includes publication-ready themes so you don’t restyle from scratch each time. Here theme_dsi() and scale_y_dsi_col() (which removes the gap between bars and the axis) do the work:

library(ggplot2)

who_countries |>
  count(who_region) |>
  ggplot(aes(reorder(who_region, n), n)) +
  geom_col(fill = "#0093D5") +
  coord_flip() +
  scale_y_dsi_col() +
  theme_dsi() +
  labs(title = "WHO Member States by region", x = NULL, y = NULL)

For multi-panel layouts there’s theme_dsi_facet(), and for tables dsi_flextable_defaults() sets booktabs styling, bold headers and sensible padding in one line. The point is that a chart pulled straight from who_countries or a cleaned indicator already carries a consistent WHO look, without per-plot fiddling.

Try it

DSIR is on CRAN and the source is on GitHub:

install.packages("DSIR")

• Documentation and a fuller walkthrough: https://shanlong-who.github.io/DSIR/
• Source, issues, and feature requests: https://github.com/shanlong-who/DSIR/

It’s early days and the package is small by design — it does the unglamorous data-plumbing so the analysis can start sooner. If you work with WHO or SDG data in R and find it useful, I’d love to hear what’s missing; issues and suggestions are very welcome. And if it saves you some glue code, a GitHub star helps other people working on global health find it too.