Web Scraping Academia Institute’s Grant Fundings using R

blog
Web Scrapping
How to use R for web scrapping
Author

Jihong Zhang

Published

May 12, 2022

This is an example of how to web scrape grants of active research in college’s official website. Please follow the academia institute’s website robots rules.

⌘+C
library(XML)
library(stringr)
library(rvest)
library(tidyverse)
library(kableExtra)

1 Web Scrapping

⌘+C
topics = rep(NA, 84)
fundings = rep(NA, 84)
iter = 0
for (page in 1:9) {
  ## parent webpage
  scrape_url <- paste0('http://XXXXXXXXXXXXXXXXXXXXXX', page)
  
  html_form_page <- read_html(scrape_url)
  
  ## find the child web page containing projects' name, total award, topic etc.
  child_url = html_form_page |> html_elements("h3 a[href]") |> html_attr("href") 
  
  for (item in 1:length(child_url)) {
    iter = iter + 1
    child_html_text <- child_url[item] |> read_html() |> html_elements("div[class='study_wrapper']") |> html_text() 
    topic = child_html_text |> str_extract(pattern = "Topic\\(s\\)\\: [a-zA-Z]+\\b") |> str_replace(pattern = "Topic\\(s\\)\\: ", "")
    funding = child_html_text |> str_extract(pattern = "\\$\\d+\\,\\d+") |> str_replace_all(pattern = "\\$|\\,", "") |> as.numeric()
    topics[iter] = topic
    fundings[iter] = funding
  }
}

dat <- data.frame(topic = topics, funding_amount = fundings) |> 
  add_row(topic = "Marijuana", funding_amount = 3743) |> 
  mutate(topic = ifelse(topic == "TobaccoMarijuana", "Tobacco", topic))
topic funding_amount
Other 97666
Other 15000
Tobacco 1271
Cancer NA
Other 199836
Other 64614
Cancer 200000
Other 75000
Other 72972
Cancer 329234

2 Visualization in ggplot2

⌘+C
## funding per project
dat1 <- dat |> 
  group_by(topic) |> 
  summarise(funding_amount_mean = mean(funding_amount, na.rm = T)) |> 
  mutate(topic = fct_reorder(topic, desc(funding_amount_mean)))

ggplot(dat1) +
  aes(x = topic, y = funding_amount_mean) +
  geom_col(fill = "darkblue") +
  scale_y_continuous(labels = scales::unit_format(unit = "M", scale = 1e-6)) +
  labs(y = "funding per project", title = "Funding for each project during 2019 to 2022") +
  theme(legend.position = "none", text = element_text(size = 12)) # remove lengend

⌘+C
## Total funding amount
dat2 <- dat |> 
  group_by(topic) |> 
  summarise(
    funding_amount_sum = sum(funding_amount, na.rm = T), 
    n = n()) |> 
  mutate(
    topic = fct_reorder(topic, desc(funding_amount_sum)),
    highlight = ifelse(funding_amount_sum == max(funding_amount_sum), 1, 0) |> as.factor())

ggplot(dat2) +
  aes(x = topic, y = funding_amount_sum, fill = highlight) +
  geom_col() +
  geom_text(aes(label = round(funding_amount_sum/ 10^6, 3)), vjust = 0.001, size = 5) +
  geom_label(aes(label = n), vjust = 0.999, size = 5, color = "white") +
  scale_fill_manual(values = c("darkblue", "red2")) +
  labs(x = "", y = "total amount of funding", 
       title = "Total Amount of Funding and Number of Grant Projects during 2019-2022", 
       subtitle = "Active research at Health Promotion Center from 2019 to 2022",
       caption = "source: https://healthpromotionresearch.org/Active-Studies/") +
  scale_y_continuous(labels = scales::unit_format(unit = "M", scale = 1e-6)) +
  theme(legend.position = "none", text = element_text(size = 10)) # remove lengend

Back to top