Clinical medicine is by far the largest field in the 2026 Stanford–Elsevier top 2% list: 76,155 of the 249,310 listed authors, 31%, have it as their main field, more than the next three of the 20 fields combined. I wanted to know where those scientists work, and whether the country ranking still looks the same once you ask how many people each country has.

The data is John Ioannidis’s science-wide author database, built with Elsevier from Scopus (Ioannidis, J.P.A. (2026), “August 2026 data-update for ‘Updated science-wide author databases of standardized citation indicators’”, Mendeley Data, v9, doi:10.17632/btchxktzyw.9, licensed CC BY-NC 3.0). An author is listed if they rank in the top 100,000 by a composite citation score or in the top 2% of their main subfield. I use the single-year table, which scores citations received during 2025. Each row carries a main field from the Science-Metrix classification (sm-field) and the country of the author’s most recent institution (cntry).

Find a top-cited scientist or institution: If you would rather skip the tutorial and find your own name or institution, topcitedscientists.com has the whole 2026 list: search for a scientist by name or see institutions ranked by their number of top-cited scientists.

How do I count the Stanford–Elsevier top 2% scientists by country in R?

Download the single-year author table from Mendeley Data, read its Data sheet with readxl, fill the blank countries from the top_cntry_career column, then filter on sm-field and count() the countries. This block runs on its own with readxl 1.5.0 and tidyverse 2.0.0 on R 4.6.1. The file is about 100 MB, so it raises the download timeout first:

library(tidyverse)
library(readxl)

url <- "https://data.mendeley.com/public-files/datasets/btchxktzyw/files/a517a94c-3fb0-4cee-b97f-f2d476df66f0/file_downloaded"
file <- "Table_1_Authors_singleyr_2025_pubs_since_1788_wopp_extracted_202608.xlsx"

options(timeout = 600)
if (!file.exists(file)) {
  download.file(url, file, mode = "wb")
}

authors <- read_excel(file, sheet = "Data") |>
  mutate(
    filled = is.na(cntry),
    cntry = coalesce(cntry, str_to_lower(top_cntry_career))
  )

clinical <- authors |>
  filter(`sm-field` == "Clinical Medicine") |>
  count(cntry, sort = TRUE) |>
  mutate(share = n / sum(n))

clinical
## # A tibble: 147 × 3
##    cntry     n  share
##    <chr> <int>  <dbl>
##  1 usa   28458 0.374 
##  2 gbr    6303 0.0828
##  3 chn    3924 0.0515
##  4 deu    3770 0.0495
##  5 ita    3710 0.0487
##  6 can    3140 0.0412
##  7 aus    2943 0.0386
##  8 fra    2120 0.0278
##  9 jpn    2074 0.0272
## 10 nld    1955 0.0257
## # ℹ 137 more rows

Three details in that code matter. download.file() gives up after 60 seconds by default, which a slow connection hits long before 100 MB. The workbook’s first sheet is a column key, so read_excel(file) without sheet = quietly returns its 53 rows of definitions instead of the data. And 169 authors have no cntry, 77 of them in clinical medicine, the most of any field; the dataset’s top_cntry_career column (the country that appears most often on their papers) fills them, and the filled column keeps a record of which rows were changed. That moves the US from 28,424 to 28,458.

The United States has 28,458 of the field’s top-cited scientists, 37.4%. The United Kingdom follows with 6,303, then China (3,924), Germany (3,770) and Italy (3,710). Those five hold 61% of the field, and 146 countries have at least one listed author.

Related posts on DataScience+:

• Machine Learning Results in R: one plot to rule them all! (Part 1 – Classification Models)
• Visualize your Portfolio’s Performance and Generate a Nice Report with R
• Creating Slopegraphs with R

Per resident, Switzerland and Denmark lead

A raw count mostly measures how big a country’s research system is. Dividing by population asks a different question: how dense the country’s pool of highly cited clinicians and medical researchers is. I take 2024 population from the World Bank’s API (indicator SP.POP.TOTL, keyless JSON) and keep countries with at least a million residents:

wb <- jsonlite::fromJSON(paste0(
  "https://api.worldbank.org/v2/country/all/indicator/",
  "SP.POP.TOTL?date=2024&format=json&per_page=400"
))

population <- wb[[2]] |>
  as_tibble() |>
  transmute(
    cntry = str_to_lower(countryiso3code),
    country = country$value,
    pop = value
  ) |>
  filter(cntry != "")

per_million <- clinical |>
  inner_join(population, by = "cntry") |>
  filter(pop >= 1e6) |>
  mutate(per_million = n / pop * 1e6) |>
  arrange(desc(per_million))

per_million |> select(country, n, per_million) |> head(8)
## # A tibble: 8 × 3
##   country            n per_million
##   <chr>          <int>       <dbl>
## 1 Switzerland     1345       149. 
## 2 Denmark          820       137. 
## 3 Netherlands     1955       109. 
## 4 Australia       2943       108. 
## 5 Sweden          1028        97.3
## 6 United Kingdom  6303        91.0
## 7 United States  28458        83.7
## 8 Canada          3140        76.1

Switzerland has 149 top-cited clinical medicine scientists per million residents and Denmark 137, against 84 in the United States, which ranks 7th per resident behind the Netherlands, Australia, Sweden and the United Kingdom. China is 2.8 per million and India 0.9. The floor matters: without it, Monaco’s three listed authors would rank it among the leaders. Taiwan (428 authors) drops out of this table because the World Bank does not publish its population.

A country’s count also reflects its field mix

China is second in the whole list, with 37,204 listed authors across all fields, yet only third in clinical medicine. The reason is what its top-cited scientists work on:

field_mix <- authors |>
  summarise(
    listed = n(),
    clinical = mean(`sm-field` == "Clinical Medicine"),
    .by = cntry
  ) |>
  arrange(desc(listed))

field_mix |> head(5)
## # A tibble: 5 × 3
##   cntry listed clinical
##   <chr>  <int>    <dbl>
## 1 usa    73722    0.386
## 2 chn    37204    0.105
## 3 gbr    18320    0.344
## 4 deu    10942    0.345
## 5 aus     8974    0.328

Clinical medicine is the main field of 39% of listed US authors and 34% of British ones, against 31% across the whole list and 11% in China, where the largest field is Enabling & Strategic Technologies (8,807 authors). So a field-by-country ranking says as much about where a country’s research effort goes as about how strong it is in that field.

Two caveats travel with every number here. The list counts citations, not quality, and the dataset’s authors warn against reading it as absolute evidence of excellence. And cntry is where an author works now, not their nationality or where they trained.

This article was first published on DataScience+, a community of R and Python tutorial authors. Have a data-science technique worth sharing? Write for us — no pitch required.