Scraping Table With R Datasets

It’s a very quick post on how to get a list of datasets available from within R with their basic description (what package they can be found in, number of observations and variables). It always takes me some time to find the right dataset to showcase whatever process or method I’m working with, so this was really to make my life easier. So! I’m going to scrape the table with a list of R datasets from here using rvest and xml2 packages:

library(rvest)
library(xml2)
library(dplyr)

url <- "https://vincentarelbundock.github.io/Rdatasets/datasets.html"

r_datasets <- read_html(url) %>% # read url
    html_nodes("table") %>% # extract all the tables
   .[[2]] %>% # it's the second table we want
    html_table() # convert it to a usable format (data.frame)

As a result, we get a tidy data frame…

str(r_datasets)

## 'data.frame':    1072 obs. of  7 variables:
##  $ Package: chr  "datasets" "datasets" "datasets" "datasets" ...
##  $ Item   : chr  "AirPassengers" "BJsales" "BOD" "CO2" ...
##  $ Title  : chr  "Monthly Airline Passenger Numbers 1949-1960" "Sales Data with Leading Indicator" "Biochemical Oxygen Demand" "Carbon Dioxide Uptake in Grass Plants" ...
##  $ Rows   : int  144 150 6 237 6 4 72 84 98 50 ...
##  $ Cols   : int  2 2 2 2 2 4 2 2 2 5 ...
##  $ csv    : chr  "CSV" "CSV" "CSV" "CSV" ...
##  $ doc    : chr  "DOC" "DOC" "DOC" "DOC" ...

library(knitr)
r_datasets %>% 
  select(-c(csv, doc)) %>% 
  head() %>%
  kable()

.. that we can filter freely, according to our needs:

r_datasets %>% filter(Rows >= 1000 & Cols >= 50) %>% kable()

r_datasets %>% filter(grepl("cat", Item)) %>% kable()

This totally maked my life easier, so hope it will help you, too!