NSF Grant on Bayesian Cognitive Diagnosis

Data visualization using ggplot2

Author

Jihong

Published

May 20, 2022

The NSF grant awards can be easily searched via NSF website. Check Active Awards and type “Bayesian & Cognitive Diagnosis” into the search box. Two things I’m really interested in: (1) Who as Principal Investigator (PI) obtained most NSF awards? (2) Which institute obtained most NSF awards?

⌘+C 
library(tidyverse)
⌘+C 
awards <- read.csv("Awards.csv")
awards <- awards |>
  filter(NSFOrganization == "SES") # select social economic sciences

1 Who get most awards

⌘+C 
pi_amount <- awards |> 
  select(PI = PrincipalInvestigator, Org = Organization, Amount = AwardedAmountToDate, Title) |> 
  filter(PI!="") |> 
  mutate(Org = str_replace_all(Org, pattern = " -DO NOT USE", "")) |> 
  mutate(Amount = str_replace_all(Amount, pattern = "[\\$\\,]", "")) |> 
  mutate(Amount = as.numeric(Amount)) |> 
  mutate(Title = str_replace(Title, "    ", " ")) |> 
  mutate(Title = str_replace(Title, "NCRN-MN: ", "")) |> 
  mutate(Title = str_replace(Title, "Collaborative Research: ", "")) |> 
  mutate(Title = str_replace(Title, "CAREER: ", "")) |> 
  mutate(Title = sub(x = Title, pattern = '(?<=.{85})', replacement = '\n', perl = TRUE)) |> 
  mutate(Title = str_replace(Title, "^ ", "")) |> 
  mutate(PI = paste0(PI, "\n", Org))

pi_amount_p <- pi_amount |> 
  group_by(PI) |> 
  summarise(
    Amount_sum = sum(Amount),
    Award_num = n(),
    Title = Title[1]
  ) |> 
  mutate(PI = fct_reorder(PI, Amount_sum)) |> 
  arrange(desc(PI)) |> 
  head(50) |>  # top 100 
  mutate(Highlight = c(rep(1, 10), rep(0, 40)),
         rank = row_number()) |> 
  mutate(labels = paste0(round(Amount_sum/ 1e6, 1), " M"))

ggplot(pi_amount_p) +
  geom_col(aes(x = PI, y = Amount_sum, fill = factor(Highlight))) +
  # geom_text(aes(x = PI, y = 0, label = labels), hjust = 1, size = 3) +
  geom_text(aes(x = PI, y = 0, label = Title), hjust = 0, size = 3, alpha = 0.8) +
  scale_y_continuous(labels = scales::label_number(suffix = " M", scale = 1e-6), limits = c(NA, 4500000)) +
  # scale_x_discrete(labels = PI, breaks = PI) +
  scale_fill_manual(values = c("darkblue", "red")) +
  labs(title = "NSF Social Economic Sciences: Top 50 PIs most amount of awards\nfunded (2021-2022)", 
       subtitle = "Bayesian Cognitive Diagnosis",
       caption = 'SEARCH: "Bayesian Cognitive Diagnosis"\nSource:https://nsf.gov/awardsearch',
       x = "PI & Orgnization", y = "Total Amount of Awards") +
  coord_flip() +
  ggdark::dark_theme_gray() +
  theme(legend.position = "", text = element_text(size = 10), title = element_text(hjust = -0.1))

Back to top