Four Questions About the Nobel Prize

Working with JSON from the Nobel Prize API in R

Author

Kevin Villa

Published

November 3, 2026

The Nobel Prize organization publishes its data through a free public API (version 2.1) at https://api.nobelprize.org/2.1. This report uses both of its main endpoints:

No API key is needed.

Getting the data

library(httr2)
library(jsonlite)
library(dplyr)
library(tidyr)
library(purrr)
library(ggplot2)

Both endpoints return at most 1,000 records per request, so this function keeps requesting the next page until a page comes back short.

get_all <- function(endpoint) {
  records <- list()
  offset <- 0
  repeat {
    page <- request(paste0("https://api.nobelprize.org/2.1/", endpoint)) |>
      req_url_query(offset = offset, limit = 1000) |>
      req_perform() |>
      resp_body_json()
    records <- c(records, page[[endpoint]])
    if (length(page[[endpoint]]) < 1000) break
    offset <- offset + 1000
  }
  records
}

laureates_raw <- get_all("laureates")
prizes_raw    <- get_all("nobelPrizes")

length(laureates_raw)
[1] 1028
length(prizes_raw)
[1] 688

Turning the JSON into tidy data frames

The JSON is deeply nested. A laureate holds a list of prizes, each prize holds a list of affiliations, and most text fields hold one value per language (en, se, no). I pulled out the English values and built three tidy tables.

laureates <- tibble(
  id            = map_chr(laureates_raw, "id"),
  type          = if_else(map_lgl(laureates_raw, \(x) is.null(x$orgName)),
                          "person", "organization"),
  name          = map_chr(laureates_raw, \(x) x$knownName$en %||% x$orgName$en),
  gender        = map_chr(laureates_raw, "gender", .default = NA_character_),
  birth_date    = map_chr(laureates_raw, list("birth", "date"), .default = NA_character_),
  birth_country = map_chr(laureates_raw, list("birth", "place", "countryNow", "en"),
                          .default = NA_character_)
) |>
  mutate(
    birth_year = as.integer(substr(birth_date, 1, 4)),
    birth_country = case_when(
      birth_country %in% c("Scotland", "Northern Ireland") ~ "United Kingdom",
      birth_country == "Faroe Islands (Denmark)" ~ "Denmark",
      birth_country == "Guadeloupe, France" ~ "France",
      .default = birth_country
    )
  )

awards <- map(laureates_raw, \(x) {
  map(x$nobelPrizes, \(p) tibble(
    id                  = x$id,
    award_year          = as.integer(p$awardYear),
    category            = p$category$en,
    portion             = p$portion,
    prize_status        = p$prizeStatus,
    affiliation         = pluck(p, "affiliations", 1, "nameNow", "en",
                                .default = NA_character_),
    affiliation_country = pluck(p, "affiliations", 1, "countryNow", "en",
                                .default = NA_character_)
  )) |>
    list_rbind()
}) |>
  list_rbind()

prizes <- tibble(
  award_year            = as.integer(map_chr(prizes_raw, "awardYear")),
  category              = map_chr(prizes_raw, list("category", "en")),
  prize_amount          = map_dbl(prizes_raw, "prizeAmount"),
  prize_amount_adjusted = map_dbl(prizes_raw, "prizeAmountAdjusted"),
  n_laureates           = map_int(prizes_raw, \(p) length(p$laureates))
)

glimpse(laureates)
Rows: 1,028
Columns: 7
$ id            <chr> "745", "102", "779", "259", "1004", "114", "982", "981",…
$ type          <chr> "person", "person", "person", "person", "person", "perso…
$ name          <chr> "A. Michael Spence", "Aage N. Bohr", "Aaron Ciechanover"…
$ gender        <chr> "male", "male", "male", "male", "male", "male", "male", …
$ birth_date    <chr> "1943-11-07", "1922-06-19", "1947-10-01", "1926-08-11", …
$ birth_country <chr> "USA", "Denmark", "Israel", "Lithuania", NA, "Pakistan",…
$ birth_year    <int> 1943, 1922, 1947, 1926, 1948, 1926, 1961, 1976, 1939, 19…
glimpse(awards)
Rows: 1,036
Columns: 7
$ id                  <chr> "745", "102", "779", "259", "1004", "114", "982", …
$ award_year          <int> 2001, 1975, 2004, 1982, 2021, 1979, 2019, 2019, 20…
$ category            <chr> "Economic Sciences", "Physics", "Chemistry", "Chem…
$ portion             <chr> "1/3", "1/3", "1/3", "1", "1", "1/3", "1/3", "1", …
$ prize_status        <chr> "received", "received", "received", "received", "r…
$ affiliation         <chr> "Stanford University", "Niels Bohr Institute", "Te…
$ affiliation_country <chr> "USA", "Denmark", "Israel", "United Kingdom", NA, …
glimpse(prizes)
Rows: 688
Columns: 5
$ award_year            <int> 1901, 1901, 1901, 1901, 1901, 1902, 1902, 1902, …
$ category              <chr> "Chemistry", "Literature", "Peace", "Physics", "…
$ prize_amount          <dbl> 150782, 150782, 150782, 150782, 150782, 141847, …
$ prize_amount_adjusted <dbl> 10892857, 10892857, 10892857, 10892857, 10892857…
$ n_laureates           <int> 1, 1, 2, 1, 1, 1, 1, 2, 2, 1, 1, 1, 1, 3, 1, 1, …
  • laureates has one row per laureate.
  • awards has one row per laureate per prize, so people who won twice, like Marie Curie, have two rows.
  • prizes has one row per category per year.

A few cleaning choices:

  • Organizations have no gender or birth date, so those fields are NA for them.
  • Some birth dates have an unknown month and day (for example 1898-00-00), so I kept only the year.
  • Many laureates in Peace and Literature have no affiliation, so affiliation is NA for them.
  • Prizes that were never awarded, such as during the World Wars, have zero laureates.
  • Birth countries list Scotland and Northern Ireland separately, but affiliations only use “United Kingdom.” I combined them so a Scottish scientist working in Edinburgh isn’t counted as having left home. I did the same for the Faroe Islands (Denmark) and Guadeloupe (France).

Question 1: Are Nobel Prizes shared more often now than in the past?

Each prize can go to up to three laureates. Using the /nobelPrizes data, I counted the laureates on every prize that was awarded and averaged them by decade.

not_awarded <- prizes |> filter(n_laureates == 0)

sharing <- prizes |>
  filter(n_laureates > 0) |>
  mutate(decade = award_year %/% 10 * 10) |>
  group_by(decade) |>
  summarise(prizes = n(),
            avg_laureates = mean(n_laureates),
            share_split_three = mean(n_laureates == 3))

ggplot(sharing, aes(decade, avg_laureates)) +
  geom_line(color = "#2a78d6", linewidth = 1) +
  geom_point(color = "#2a78d6", size = 2.5) +
  scale_x_continuous(breaks = seq(1900, 2020, 20)) +
  labs(title = "Average number of laureates per Nobel Prize, by decade",
       x = NULL, y = "Laureates per prize") +
  theme_minimal()

sharing_by_category <- prizes |>
  filter(n_laureates > 0) |>
  mutate(period = if_else(award_year < 1950, "1901-1949", "2000 onward")) |>
  filter(award_year < 1950 | award_year >= 2000) |>
  group_by(category, period) |>
  summarise(avg_laureates = mean(n_laureates), .groups = "drop") |>
  pivot_wider(names_from = period, values_from = avg_laureates)

knitr::kable(sharing_by_category, digits = 2)
category 1901-1949 2000 onward
Chemistry 1.20 2.59
Economic Sciences NA 2.11
Literature 1.05 1.00
Peace 1.37 1.41
Physics 1.23 2.67
Physiology or Medicine 1.40 2.44 
first_dec <- sharing |> slice_min(decade, n = 1)
last_dec  <- sharing |> slice_max(decade, n = 1)

Answer: Yes. In the 1900s, a prize went to 1.27 laureates on average. In the 2020s, the average is 2.05, and 40% of prizes were split three ways. The table shows that the change comes from the science prizes. Literature has almost always gone to a single winner. Separately, 49 prizes were never awarded at all, 31 of them during the two World Wars.

Question 2: How old are laureates when they win, and has that changed?

This joins each award to the laureate’s birth year and subtracts. Because some birth dates are missing the month and day, ages can be off by up to one year.

ages <- awards |>
  inner_join(laureates |> filter(type == "person"), by = "id") |>
  filter(!is.na(birth_year)) |>
  mutate(age = award_year - birth_year,
         era = if_else(award_year < 1960, "1901-1959", "1960 onward"))

age_by_category <- ages |>
  group_by(category, era) |>
  summarise(median_age = median(age), .groups = "drop") |>
  pivot_wider(names_from = era, values_from = median_age)

knitr::kable(age_by_category)
category 1901-1959 1960 onward
Chemistry 50 63.0
Economic Sciences NA 67.0
Literature 60 69.0
Peace 68 59.0
Physics 47 61.5
Physiology or Medicine 54 61.0 
ggplot(ages, aes(award_year, age)) +
  geom_point(alpha = 0.3, color = "#2a78d6") +
  geom_smooth(method = "loess", se = FALSE, color = "#eb6834") +
  facet_wrap(~category) +
  labs(title = "Age of laureates at the time of their award",
       x = "Award year", y = "Age") +
  theme_minimal()

youngest <- ages |> slice_min(age, n = 1, with_ties = FALSE)
oldest   <- ages |> slice_max(age, n = 1, with_ties = FALSE)

went_down <- age_by_category |>
  filter(`1960 onward` < `1901-1959`) |>
  pull(category)

Answer: The median age at the time of the award is 61. Malala Yousafzai is the youngest laureate, winning the Peace prize in 2014 at about 17. John B. Goodenough is the oldest, winning the Chemistry prize in 2019 at about 97. The table compares each category before and after 1960. The median age went up in every category except Peace.

Question 3: When did each category get its first woman, and how has that changed?

This filters the awards to people, joins on gender, and finds the earliest award to a woman in each category.

person_awards <- awards |>
  inner_join(laureates |> filter(type == "person"), by = "id")

first_women <- person_awards |>
  filter(gender == "female") |>
  group_by(category) |>
  slice_min(award_year, n = 1, with_ties = FALSE) |>
  ungroup() |>
  select(category, award_year, name) |>
  arrange(award_year)

knitr::kable(first_women, col.names = c("Category", "Year", "First woman laureate"))
Category Year First woman laureate
Physics 1903 Marie Curie
Peace 1905 Bertha von Suttner
Literature 1909 Selma Lagerlöf
Chemistry 1911 Marie Curie
Physiology or Medicine 1947 Gerty Cori
Economic Sciences 2009 Elinor Ostrom 
women_by_decade <- person_awards |>
  mutate(decade = award_year %/% 10 * 10) |>
  group_by(decade) |>
  summarise(share_women = mean(gender == "female"))

ggplot(women_by_decade, aes(decade, share_women)) +
  geom_col(fill = "#2a78d6") +
  scale_y_continuous(labels = scales::percent) +
  scale_x_continuous(breaks = seq(1900, 2020, 20)) +
  labs(title = "Share of Nobel Prizes awarded to women, by decade",
       x = NULL, y = NULL) +
  theme_minimal()

recent <- women_by_decade |> slice_max(decade, n = 1)

Answer: Marie Curie was the first woman to win, in Physics in 1903. The last category to award a woman was Economic Sciences, in 2009. Overall, women have received 7.0% of all prizes awarded to people. The share is rising, reaching 20% in the 2020s.

Question 4: Which countries lose the most laureates to other countries?

This compares two fields for every laureate: the country they were born in, and the country of the institution they were working at when they won. Both use today’s country names, so they can be compared directly. A laureate counts as having left if the two countries are different. I counted each person once, using their first prize, and left out laureates with no affiliation (mostly Peace and Literature).

moves <- awards |>
  inner_join(laureates |> filter(type == "person"), by = "id") |>
  filter(!is.na(birth_country), !is.na(affiliation_country)) |>
  group_by(id) |>
  slice_min(award_year, n = 1, with_ties = FALSE) |>
  ungroup() |>
  mutate(left_home = birth_country != affiliation_country)

lost <- moves |>
  filter(left_home) |>
  count(birth_country, name = "laureates_lost", sort = TRUE)

gained <- moves |>
  filter(left_home) |>
  count(affiliation_country, name = "laureates_gained", sort = TRUE)

knitr::kable(head(lost, 10), col.names = c("Born in", "Laureates who won while working elsewhere"))
Born in Laureates who won while working elsewhere
United Kingdom 26
Germany 24
Poland 19
Canada 15
France 13
Russia 11
Austria 10
the Netherlands 9
Hungary 8
Italy 8 
knitr::kable(head(gained, 10), col.names = c("Working in", "Laureates born elsewhere"))
Working in Laureates born elsewhere
USA 134
United Kingdom 28
Germany 25
Switzerland 12
France 11
Canada 6
Sweden 4
Russia 3
Australia 2
Austria 2 
top_lost <- lost |> slice_head(n = 10)
where_they_went <- moves |>
  filter(left_home, birth_country %in% top_lost$birth_country) |>
  count(birth_country, affiliation_country) |>
  group_by(birth_country) |>
  slice_max(n, n = 1, with_ties = FALSE) |>
  ungroup()

ggplot(top_lost, aes(laureates_lost, reorder(birth_country, laureates_lost))) +
  geom_col(fill = "#2a78d6", width = 0.7) +
  labs(title = "Countries that lost the most future laureates",
       subtitle = "Born there, but working in another country when they won",
       x = "Laureates", y = NULL) +
  theme_minimal()

knitr::kable(where_they_went,
             col.names = c("Born in", "Most common destination", "Laureates"))
Born in Most common destination Laureates
Austria Germany 4
Canada USA 15
France USA 8
Germany USA 11
Hungary USA 4
Italy USA 5
Poland Germany 9
Russia USA 5
United Kingdom USA 20
the Netherlands USA 4

Answer: Of the 757 laureates with both a birth country and an affiliation, 236 (31%) won while working outside the country where they were born. United Kingdom lost the most, with 26 laureates, followed by Germany with 24. USA gained by far the most, with 134 laureates who were born somewhere else. Two things to keep in mind. First, the API records where laureates worked, not their citizenship, so this measures where they did their prize-winning work rather than which passport they held. Second, birth countries use today’s borders. Someone born in Breslau when it was part of Germany counts as born in Poland, which is part of why Poland ranks high.

Source

Data from the Nobel Prize API v2.1 (https://api.nobelprize.org/2.1), published by Nobel Prize Outreach under its terms of use.