Show code
library(httr2)
library(jsonlite)
library(dplyr)
library(tidyr)
library(stringr)
library(ggplot2)
library(tidytext)
library(knitr)
library(scales)
theme_set(theme_minimal(base_size = 12))| Data | Nobel Prize API v2.1, /nobelPrizes endpoint, pulled as JSON |
| Focus | The prize itself, not the winners: when it wasn’t given, what it was worth, how many people shared it, and what it was given for |
| Join/compare question | Q1 compares every year and category the prize could have been given against the ones it was |
library(httr2)
library(jsonlite)
library(dplyr)
library(tidyr)
library(stringr)
library(ggplot2)
library(tidytext)
library(knitr)
library(scales)
theme_set(theme_minimal(base_size = 12))| Endpoint | https://api.nobelprize.org/2.1/nobelPrizes |
| Key needed | No |
| Paging | offset and limit. The first response’s meta$count gives the total, so the code loops until it has every prize |
| Sort | asc, oldest first |
get_page <- function(offset, limit = 100) {
request("https://api.nobelprize.org/2.1/nobelPrizes") |>
req_url_query(offset = offset, limit = limit, sort = "asc") |>
req_retry(max_tries = 3) |>
req_perform() |>
resp_body_json()
}
fetch_all_prizes <- function() {
first_page <- get_page(0)
total <- first_page$meta$count
other_pages <- lapply(seq(100, total - 1, by = 100), get_page)
pages <- c(list(first_page), other_pages)
do.call(c, lapply(pages, `[[`, "nobelPrizes"))
}
snapshot_file <- "nobel_prizes_raw.json"
prizes_raw <- tryCatch(
fetch_all_prizes(),
error = function(e) {
message("API unavailable, using saved snapshot: ", conditionMessage(e))
read_json(snapshot_file)
}
)
write_json(prizes_raw, snapshot_file, auto_unbox = TRUE, digits = NA)
length(prizes_raw)[1] 688
The live API is tried first. Every successful pull also saves the raw JSON to nobel_prizes_raw.json, so the report still renders if the API is down.
Each prize is a nested list. Category names come in three languages, and the laureates are a list inside the prize:
str(prizes_raw[[1]], max.level = 2, list.len = 10)List of 8
$ awardYear : chr "1901"
$ category :List of 3
..$ en: chr "Chemistry"
..$ no: chr "Kjemi"
..$ se: chr "Kemi"
$ categoryFullName :List of 3
..$ en: chr "The Nobel Prize in Chemistry"
..$ no: chr "Nobelprisen i kjemi"
..$ se: chr "Nobelpriset i kemi"
$ dateAwarded : chr "1901-11-12"
$ prizeAmount : int 150782
$ prizeAmountAdjusted: int 10892857
$ links :List of 1
..$ :List of 4
$ laureates :List of 1
..$ :List of 7
Two tidy tables, built with tidyr’s rectangling functions:
| Table | One row per | Built with |
|---|---|---|
prizes |
Prize (year + category) | hoist() pulls fields out of each prize, including nested ones like category$en |
prize_laureates |
Laureate per prize | unnest_longer() turns each prize’s laureate list into rows, then hoist() pulls their fields |
sek_per_usd <- 9.89 # Swedish kronor per US dollar, early November 2026 (see Data Source)
usd <- function(x) vapply(x, function(v) dollar(v, accuracy = if (v < 1e6) 1 else 0.1,
scale_cut = cut_short_scale()), "")
prizes <- tibble(prize = prizes_raw) |>
hoist(prize,
year = "awardYear",
category = c("category", "en"),
amount = "prizeAmount",
amount_adjusted = "prizeAmountAdjusted",
top_motivation = c("topMotivation", "en"),
laureates = "laureates") |>
select(-prize) |>
mutate(year = as.integer(year),
n_laureates = lengths(laureates),
value_usd = amount_adjusted / sek_per_usd)
prize_laureates <- prizes |>
select(year, category, value_usd, laureates) |>
unnest_longer(laureates) |>
hoist(laureates,
laureate_id = "id",
name = c("knownName", "en"),
org_name = c("orgName", "en"),
portion = "portion",
motivation = c("motivation", "en")) |>
select(-laureates) |>
mutate(name = coalesce(name, org_name),
share = 1 / as.numeric(str_extract(portion, "\\d+$")),
take = value_usd * share)
prizes |>
head(5) |>
transmute(Year = year, Category = category,
`Paid then (SEK)` = comma(amount),
`Worth today (SEK)` = comma(amount_adjusted),
`Worth today (USD)` = dollar(value_usd, accuracy = 1),
Winners = n_laureates)| Year | Category | Paid then (SEK) | Worth today (SEK) | Worth today (USD) | Winners |
|---|---|---|---|---|---|
| 1901 | Chemistry | 150,782 | 10,892,857 | $1,101,401 | 1 |
| 1901 | Literature | 150,782 | 10,892,857 | $1,101,401 | 1 |
| 1901 | Peace | 150,782 | 10,892,857 | $1,101,401 | 2 |
| 1901 | Physics | 150,782 | 10,892,857 | $1,101,401 | 1 |
| 1901 | Physiology or Medicine | 150,782 | 10,892,857 | $1,101,401 | 1 |
prize_laureates |> select(year, category, name, portion, share, motivation) |> head(5)| year | category | name | portion | share | motivation |
|---|---|---|---|---|---|
| 1901 | Chemistry | Jacobus H. van ’t Hoff | 1 | 1.0 | in recognition of the extraordinary services he has rendered by the discovery of the laws of chemical dynamics and osmotic pressure in solutions |
| 1901 | Literature | Sully Prudhomme | 1 | 1.0 | in special recognition of his poetic composition, which gives evidence of lofty idealism, artistic perfection and a rare combination of the qualities of both heart and intellect |
| 1901 | Peace | Henry Dunant | 1/2 | 0.5 | for his humanitarian efforts to help wounded soldiers and create international understanding |
| 1901 | Peace | Frédéric Passy | 1/2 | 0.5 | for his lifelong work for international peace conferences, diplomacy and arbitration |
| 1901 | Physics | Wilhelm Conrad Röntgen | 1 | 1.0 | in recognition of the extraordinary services he has rendered by the discovery of the remarkable rays subsequently named after him |
Checks before answering anything:
share_check <- prize_laureates |>
group_by(year, category) |>
summarise(total_share = sum(share), .groups = "drop")
tibble(
Check = c("Prizes pulled", "Laureate rows",
"Prizes with no laureates", "Prizes whose shares add up to 1"),
Result = c(nrow(prizes), nrow(prize_laureates),
sum(prizes$n_laureates == 0),
paste(sum(abs(share_check$total_share - 1) < 1e-9), "of", nrow(share_check)))
)| Check | Result |
|---|---|
| Prizes pulled | 688 |
| Laureate rows | 1036 |
| Prizes with no laureates | 49 |
| Prizes whose shares add up to 1 | 639 of 639 |
Money: the API reports amounts in Swedish kronor (SEK), the currency the Nobel Foundation pays in. prizeAmountAdjusted is each prize in today’s kronor, adjusted for inflation, so years compare fairly. I convert it to today’s US dollars at 9.89 kronor per dollar (value_usd). prizeAmount, what was actually paid at the time, stays in kronor, since converting a 1901 payment at today’s exchange rate wouldn’t mean anything.
Which years and categories had no prize awarded, and why do they cluster?
Approach: build every year-and-category combination the prize could have been given (Economics only from 1969), then anti_join() it against the prizes that actually have laureates. Whatever is left was never awarded.
categories <- tibble(
category = c("Physics", "Chemistry", "Physiology or Medicine",
"Literature", "Peace", "Economic Sciences"),
first_year = c(rep(1901L, 5), 1969L)
)
all_possible <- categories |>
cross_join(tibble(year = 1901:max(prizes$year))) |>
filter(year >= first_year)
awarded <- prizes |> filter(n_laureates > 0)
not_awarded <- all_possible |>
anti_join(awarded, by = c("year", "category")) |>
mutate(period = case_when(
between(year, 1914, 1918) ~ "World War I (1914-1918)",
between(year, 1939, 1945) ~ "World War II (1939-1945)",
TRUE ~ "Other years"
))
missing_from_api <- all_possible |> anti_join(prizes, by = c("year", "category"))category_order <- c("Economic Sciences", "Peace", "Literature",
"Physiology or Medicine", "Chemistry", "Physics")
all_possible |>
left_join(not_awarded |> mutate(gap = TRUE) |> select(year, category, gap),
by = c("year", "category")) |>
mutate(status = if_else(is.na(gap), "Awarded", "Not awarded"),
category = factor(category, levels = category_order)) |>
ggplot(aes(year, category, fill = status)) +
annotate("rect", xmin = 1913.5, xmax = 1918.5, ymin = -Inf, ymax = Inf, fill = "#f3e3c3") +
annotate("rect", xmin = 1938.5, xmax = 1945.5, ymin = -Inf, ymax = Inf, fill = "#f3e3c3") +
geom_tile(height = 0.75) +
annotate("text", x = c(1916, 1942), y = 6.65, label = c("WWI", "WWII"),
size = 3.5, fontface = "bold", color = "#8a6d2f") +
scale_fill_manual(values = c(Awarded = "#d5dce4", `Not awarded` = "#c0392b")) +
scale_x_continuous(breaks = seq(1900, 2025, by = 25)) +
coord_cartesian(clip = "off") +
labs(x = NULL, y = NULL, fill = NULL) +
theme(legend.position = "top", panel.grid = element_blank())not_awarded |>
count(period, name = "Prizes not awarded") |>
mutate(Share = percent(`Prizes not awarded` / sum(`Prizes not awarded`), accuracy = 1)) |>
rename(Period = period)| Period | Prizes not awarded | Share |
|---|---|---|
| Other years | 18 | 37% |
| World War I (1914-1918) | 13 | 27% |
| World War II (1939-1945) | 18 | 37% |
not_awarded |>
group_by(Category = category) |>
summarise(`Not awarded` = n(),
`During a World War` = sum(period != "Other years"),
`Last time` = max(year)) |>
arrange(desc(`Not awarded`))| Category | Not awarded | During a World War | Last time |
|---|---|---|---|
| Peace | 19 | 9 | 1972 |
| Physiology or Medicine | 9 | 7 | 1942 |
| Chemistry | 8 | 5 | 1942 |
| Literature | 7 | 6 | 1943 |
| Physics | 6 | 4 | 1942 |
Answer:
How has the real value of the prize changed since 1901?
Approach: plot each year’s prize in today’s US dollars (value_usd). All categories get the same amount in a given year, so there’s one value per year.
money <- prizes |>
group_by(year) |>
summarise(nominal = median(amount),
adjusted = median(value_usd))
milestones <- bind_rows(
money |> filter(year == min(year)) |> mutate(label = "First prize", vjust = -0.5),
money |> slice_min(adjusted, n = 1) |> mutate(label = "Lowest", vjust = 1.4),
money |> slice_max(adjusted, n = 1) |> mutate(label = "Highest", vjust = -0.5),
money |> filter(year == max(year)) |> mutate(label = "Latest", vjust = -0.5)
)ggplot(money, aes(year, adjusted)) +
geom_line(color = "#2c5f8a", linewidth = 0.9) +
geom_point(data = milestones, size = 3, color = "#c0392b") +
geom_text(data = milestones,
aes(label = paste0(label, "\n", year, ": ", usd(adjusted)),
vjust = vjust),
size = 3.2, lineheight = 0.9) +
scale_y_continuous(limits = c(0, 1.95e6), labels = label_dollar(scale_cut = cut_short_scale())) +
scale_x_continuous(breaks = seq(1900, 2025, by = 25), expand = expansion(add = 7)) +
labs(x = NULL, y = "Prize value in today's US dollars")Paid vs. worth. How much bigger is the prize than in 1901? There are two ways to answer:
Both are shown as “times the 1901 prize,” so 1x means the same as 1901 and 2x means double:
growth <- money |>
mutate(`Kronor on the check` = nominal / nominal[year == min(year)],
`Buying power` = adjusted / adjusted[year == min(year)]) |>
select(year, `Kronor on the check`, `Buying power`) |>
pivot_longer(-year, names_to = "measure", values_to = "multiple")
growth_ends <- growth |> filter(year == max(year))
ggplot(growth, aes(year, multiple, color = measure)) +
geom_hline(yintercept = 1, color = "grey60", linetype = "dashed") +
geom_line(linewidth = 1) +
geom_text(data = growth_ends, aes(label = paste0(round(multiple, 1), "x")),
hjust = -0.2, size = 3.5, show.legend = FALSE) +
scale_color_manual(values = c(`Kronor on the check` = "#c0392b",
`Buying power` = "#2c5f8a")) +
scale_y_log10(breaks = c(0.25, 0.5, 1, 2, 5, 10, 25, 50, 100),
labels = function(x) paste0(x, "x")) +
scale_x_continuous(breaks = seq(1900, 2025, by = 25), expand = expansion(add = c(2, 10))) +
labs(x = NULL, y = "Times the 1901 prize (log scale)", color = NULL) +
theme(legend.position = "top")Takeaway: the check is about 80 times bigger than in 1901, but prices rose almost as much, so its buying power is only about 1.1 times the first prize. For most of the 20th century (the blue line below 1x), the prize bought far less than it did in 1901.
milestones |>
transmute(Milestone = label, Year = year,
`Amount paid at the time (SEK)` = comma(nominal),
`In today's US dollars` = dollar(adjusted, accuracy = 1),
`vs. 1901` = percent(adjusted / money$adjusted[money$year == min(money$year)], accuracy = 1))| Milestone | Year | Amount paid at the time (SEK) | In today’s US dollars | vs. 1901 |
|---|---|---|---|---|
| First prize | 1901 | 150,782 | $1,101,401 | 100% |
| Lowest | 1919 | 133,127 | $305,624 | 28% |
| Highest | 2001 | 10,000,000 | $1,580,666 | 144% |
| Latest | 2026 | 12,000,000 | $1,213,347 | 110% |
Answer:
Are prizes being shared by more winners over time, and how much does each winner get as a result?
Approach: count laureates per awarded prize by decade. Then join each laureate’s portion (1, 1/2, 1/3 or 1/4) to the prize’s adjusted amount to get what each winner actually took home.
split_by_decade <- awarded |>
mutate(decade = year %/% 10 * 10,
winners = factor(n_laureates, levels = 3:1,
labels = c("Three winners", "Two winners", "One winner"))) |>
count(decade, winners) |>
group_by(decade) |>
mutate(share = n / sum(n)) |>
ungroup()
ggplot(split_by_decade, aes(decade, share, fill = winners)) +
geom_col(width = 8) +
scale_fill_manual(values = c(`One winner` = "#c9d6e3", `Two winners` = "#6f9cc3",
`Three winners` = "#1b3a57")) +
scale_y_continuous(labels = label_percent()) +
scale_x_continuous(breaks = seq(1900, 2020, by = 20), labels = function(x) paste0(x, "s")) +
labs(x = NULL, y = "Share of prizes awarded", fill = NULL) +
theme(legend.position = "top")By category, average winners per prize then and now:
awarded |>
mutate(era = case_when(year < 1950 ~ "1901-1949",
year >= 2000 ~ "2000-today",
TRUE ~ "1950-1999")) |>
group_by(Category = category, era) |>
summarise(avg = round(mean(n_laureates), 2), .groups = "drop") |>
pivot_wider(names_from = era, values_from = avg) |>
select(Category, `1901-1949`, `1950-1999`, `2000-today`) |>
arrange(desc(`2000-today`)) |>
mutate(across(-Category, ~ if_else(is.na(.x), "Not yet created", format(.x, nsmall = 2))))| Category | 1901-1949 | 1950-1999 | 2000-today |
|---|---|---|---|
| Physics | 1.23 | 2.12 | 2.67 |
| Chemistry | 1.20 | 1.66 | 2.59 |
| Physiology or Medicine | 1.40 | 2.26 | 2.44 |
| Economic Sciences | Not yet created | 1.42 | 2.11 |
| Peace | 1.37 | 1.29 | 1.41 |
| Literature | 1.05 | 1.04 | 1.00 |
What each winner takes home, compared with the full prize:
take_by_decade <- prize_laureates |>
mutate(decade = year %/% 10 * 10) |>
group_by(decade) |>
summarise(`Full prize` = median(value_usd),
`Typical winner's share` = median(take)) |>
pivot_longer(-decade, names_to = "measure", values_to = "value")
ggplot(take_by_decade, aes(decade, value, color = measure)) +
geom_line(linewidth = 1) +
geom_point(size = 2) +
scale_color_manual(values = c(`Full prize` = "#9aa5b1", `Typical winner's share` = "#c0392b")) +
scale_y_continuous(limits = c(0, NA), labels = label_dollar(scale_cut = cut_short_scale())) +
scale_x_continuous(breaks = seq(1900, 2020, by = 20), labels = function(x) paste0(x, "s")) +
labs(x = NULL, y = "Today's US dollars, median", color = NULL) +
theme(legend.position = "top")Answer:
How has the wording of the prize motivations changed over the century?
Approach: each laureate has a motivation (“for the discovery of…”). Winners who share the same motivation are counted once, then the text is split into words with unnest_tokens(), as in 10A.
motivations <- prize_laureates |>
distinct(year, category, motivation) |>
mutate(decade = year %/% 10 * 10,
era = case_when(year < 1950 ~ "1901-1949",
year >= 2000 ~ "2000-today",
TRUE ~ "1950-1999"))
motivation_words <- motivations |>
unnest_tokens(word, motivation) |>
anti_join(stop_words, by = "word")How often key words appear, by decade:
keywords <- c(`"recognition"` = "recognition",
`"discover"` = "discover",
`"develop"` = "develop",
`"contribution"` = "contribution")
keyword_trend <- motivations |>
mutate(text = str_to_lower(motivation)) |>
cross_join(tibble(keyword = names(keywords), pattern = unname(keywords))) |>
mutate(has_word = str_detect(text, pattern)) |>
group_by(decade, keyword) |>
summarise(share = mean(has_word), .groups = "drop") |>
mutate(keyword = factor(keyword, levels = names(keywords)))
ggplot(keyword_trend, aes(decade, share, color = keyword)) +
geom_line(linewidth = 1) +
geom_point(size = 2) +
scale_color_manual(values = c("#9aa5b1", "#2c5f8a", "#c0392b", "#d4a017")) +
scale_y_continuous(labels = label_percent()) +
scale_x_continuous(breaks = seq(1900, 2020, by = 20), labels = function(x) paste0(x, "s")) +
labs(x = NULL, y = "Share of motivations containing the word", color = NULL) +
theme(legend.position = "top")Top words in each category, then and now (numbers are how many motivations use the word):
motivation_words |>
filter(era != "1950-1999") |>
count(category, era, word) |>
group_by(category, era) |>
slice_max(n, n = 5, with_ties = FALSE) |>
summarise(words = paste0(word, " (", n, ")", collapse = ", "), .groups = "drop") |>
pivot_wider(names_from = era, values_from = words, values_fill = "(category didn't exist)") |>
rename(Category = category)| Category | 1901-1949 | 2000-today |
|---|---|---|
| Chemistry | recognition (14), chemistry (12), investigations (11), discovery (9), services (9) | development (12), discovery (7), studies (6), protein (5), synthesis (5) |
| Economic Sciences | (category didn’t exist) | analysis (13), economic (8), theory (8), contributions (5), analyzing (4) |
| Literature | recognition (9), life (7), poetic (7), art (6), poetry (6) | poetic (4), power (4), contemporary (3), experience (3), human (3) |
| Peace | peace (21), international (14), role (11), nations (9), war (9) | efforts (16), peace (10), rights (10), democracy (9), human (7) |
| Physics | discovery (25), recognition (8), radiation (7), services (7), physics (6) | discovery (16), contributions (6), optical (5), light (4), pioneering (4) |
| Physiology or Medicine | discovery (23), discoveries (8), recognition (8), played (4), relating (4) | discoveries (19), discovery (10), cells (5), system (4), cell (3) |
Answer:
| Question | Answer |
|---|---|
| Q1 The empty years | 49 prizes never awarded, 63% during the World Wars. Peace is skipped the most and is the only one skipped after 1948 |
| Q2 The money | Fell to 28% of its 1901 value by 1919, recovered in 1991, peaked in 2001, now $1.2M in today’s dollars |
| Q3 Crowding | Three-way splits: 1% before 1940, 40% since 2000. The typical winner’s share is less than half what it was |
| Q4 The wording | From formal “recognition” to plain “for…”. “Development” became Chemistry’s top word; Peace moved toward human rights and democracy |
Nobel Prize Outreach. (2026). Nobel Prize API v2.1 [Data set]. Nobel Prize Developer Zone. https://www.nobelprize.org/about/developer-zone-2/. Accessed November 8, 2026. Terms of use: https://www.nobelprize.org/about/terms-of-use-for-api-nobelprize-org-and-data-nobelprize-org/
Exchange rate: USD/SEK 9.89, early November 2026. Pound Sterling Live (Barchart data). https://www.poundsterlinglive.com/data/currencies/usd-pairs/USDSEK-exchange-rate. Accessed November 8, 2026.
Claude (Anthropic) helped write and test the code in this report: the paged API requests, the rectangling of the nested JSON, the four analyses and the charts. I chose the four questions in my approach.
Anthropic. (2026). Claude [Large language model]. https://claude.ai/. Accessed November 8, 2026.