The goal of ralger is to facilitate web scraping in R. For a quick video tutorial, I gave a talk at useR2020, which you can find here
You can install the ralger package from CRAN with:
install.packages("ralger")or you can install the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("feddelegrand7/ralger")scrap()This is an example which shows how to extract top ranked universities’ names according to the ShanghaiRanking Consultancy:
library(ralger)
my_link <- "http://www.shanghairanking.com/ARWU2020.html"
my_node <- "#UniversityRanking a" # The class ID , we recommend SelectorGadget
best_uni <- scrap(link = my_link, node = my_node)
head(best_uni, 10)
#> [1] "Harvard University"
#> [2] "Stanford University"
#> [3] "University of Cambridge"
#> [4] "Massachusetts Institute of Technology (MIT)"
#> [5] "University of California, Berkeley"
#> [6] "Princeton University"
#> [7] "Columbia University"
#> [8] "California Institute of Technology"
#> [9] "University of Oxford"
#> [10] "University of Chicago"Thanks to the robotstxt, you can set askRobot = T to ask the robots.txt file if it’s permitted to scrape a specific web page.
If you want to scrap multiple list pages, just use scrap() in conjunction with paste0().
table_scrap()If you want to extract an HTML Table, you can use the table_scrap() function. Take a look at this webpage which lists the highest gross revenues in the cinema industry. You can extract the HTML table as follows:
data <- table_scrap(link ="https://www.boxofficemojo.com/chart/top_lifetime_gross/?area=XWW")
head(data)
#> Rank Title Lifetime Gross Year
#> 1 1 Avengers: Endgame $2,797,800,564 2019
#> 2 2 Avatar $2,790,439,092 2009
#> 3 3 Titanic $2,471,754,307 1997
#> 4 4 Star Wars: Episode VII - The Force Awakens $2,068,454,310 2015
#> 5 5 Avengers: Infinity War $2,048,359,754 2018
#> 6 6 Jurassic World $1,670,516,444 2015When you deal with a web page that contains many HTML table you can use the choose argument to target a specific table
tidy_scrap()Sometimes you’ll find some useful information on the internet that you want to extract in a tabular manner however these information are not provided in an HTML format. In this context, you can use the tidy_scrap() function which returns a tidy data frame according to the arguments that you introduce. The function takes four arguments:
We’ll work on the famous IMDb website. Let’s say we need a data frame composed of:
We will need to use the tidy_scrap() function as follows:
my_link <- "https://www.imdb.com/search/title/?groups=top_250&sort=user_rating"
my_nodes <- c(
".lister-item-header a", # The title
".text-muted.unbold", # The year of release
".ratings-imdb-rating strong" # The rating)
)
names <- c("title", "year", "rating") # respect the nodes order
tidy_scrap(link = my_link, nodes = my_nodes, colnames = names)
#> # A tibble: 50 x 3
#> title year rating
#> <chr> <chr> <chr>
#> 1 The Shawshank Redemption (1994) 9.3
#> 2 The Godfather (1972) 9.2
#> 3 The Dark Knight (2008) 9.0
#> 4 The Godfather: Part II (1974) 9.0
#> 5 12 Angry Men (1957) 9.0
#> 6 The Lord of the Rings: The Return of the King (2003) 8.9
#> 7 Pulp Fiction (1994) 8.9
#> 8 Schindler's List (1993) 8.9
#> 9 Inception (2010) 8.8
#> 10 Fight Club (1999) 8.8
#> # ... with 40 more rowsNote that all columns will be of character class. you’ll have to convert them according to your needs.
titles_scrap()Using titles_scrap(), one can efficiently scrape titles which correspond to the h1, h2 & h3 HTML tags.
If we go to the New York Times, we can easily extract the titles displayed within a specific web page :
titles_scrap(link = "https://www.nytimes.com/")
#> [1] "Listen to ‘The Daily’"
#> [2] "Introducing ‘The Argument’ With Jane Coaston"
#> [3] "DealBook D.C. Policy Project"
#> [4] "NASA Prepares to Land Perseverance Rover on Mars"
#> [5] "Immigration Overhaul Would Offer 8-Year Path to Citizenship for Millions"
#> [6] "Biden’s Plan to Link Arms With Europe Against Russia and China Isn’t So Simple"
#> [7] "How a Hardcore Liberal Lawyer Joined the Pro-Trump Mob"
#> [8] "Grilled in Hearing, Robinhood Head Apologizes for Limiting GameStop Trades"
#> [9] "Work Will Never Be the Same"
#> [10] "As Israel Reopens, ‘Whoever Does Not Get Vaccinated Will Be Left Behind’"
#> [11] "Women Leaving Work Force in Pandemic Is ‘National Emergency,’ Harris Says"
#> [12] "After getting an early start to inoculations, Canadians are worrying as the country falls behind."
#> [13] "A College Program for Disadvantaged Teens Could Shake Up Elite Admissions"
#> [14] "Pigeon Guys Face Tough Times: ‘Who Has the Money? Who Has the Roof?’"
#> [15] "Texas Could Have Kept the Lights On"
#> [16] "The Government Has Not Explained How These 13 People Were Selected to Die"
#> [17] "‘There’s No Natural Dignity in Work’"
#> [18] "Don’t Go Down the Rabbit Hole"
#> [19] "A Plan to Future-Proof the Texas Power Grid"
#> [20] "The Real Story of the ‘Draft Riots’"
#> [21] "The Romney Family Plan Sees the True Value of Parenting"
#> [22] "No Parties. No Sports. How Oberlin College Is Surviving the Pandemic."
#> [23] "Trump’s Dreaded Nickname"
#> [24] "These Women Confront Dictators. Why Can’t We?"
#> [25] "Where Do the Rohingya Go After the Coup in Myanmar?"
#> [26] "In L.A., a Health Food Store Has Become the Place to See and Be Seen"
#> [27] "These Two LeBron James Cards Are Worth $7 Million"
#> [28] "Their Upper East Side Rental Was ‘Nearly Perfect.’ Until It Wasn’t."
#> [29] "Site Index"
#> [30] "Site Information Navigation"
#> [31] "As Power Begins to Return, Many Texans Lack Drinking Water"
#> [32] "Texas Storms, California Heat Waves and ‘Vulnerable’ Utilities"
#> [33] "Senator Ted Cruz left for Cancún as Texas was battered by a storm, sparking an outcry. He planned to return Thursday."
#> [34] "About 3,500 threatened sea turtles were rescued from frigid water in Texas."
#> [35] "Opinion"
#> [36] "Editors’ Picks"
#> [37] "Advertisement"Further, it’s possible to filter the results using the contain argument:
titles_scrap(link = "https://www.nytimes.com/", contain = "TrUMp", case_sensitive = FALSE)
#> [1] "How a Hardcore Liberal Lawyer Joined the Pro-Trump Mob"
#> [2] "Trump’s Dreaded Nickname"paragraphs_scrap()In the same way, we can use the paragraphs_scrap() function to extract paragraphs. This function relies on the p HTML tag.
Let’s get some paragraphs from the lovely ropensci.org website:
paragraphs_scrap(link = "https://ropensci.org/")
#> [1] ""
#> [2] "We help develop R packages for the sciences via community driven learning, review and\nmaintenance of contributed software in the R ecosystem"
#> [3] "Use our carefully vetted, staff- and community-contributed R software tools that lower barriers to working with local and remote scientific data sources. Combine our tools with the rich ecosystem of R packages."
#> [4] "Workflow Tools for Your Code and Data"
#> [5] "Get Data from the Web"
#> [6] "Convert and Munge Data"
#> [7] "Document and Release Your Data"
#> [8] "Visualize Data"
#> [9] "Work with Databases From R"
#> [10] "Access, Manipulate, Convert Geospatial Data"
#> [11] "Interact with Web Resources"
#> [12] "Use Image & Audio Data"
#> [13] "Analyze Scientific Papers (and Text in General)"
#> [14] "Secure Your Data and Workflow"
#> [15] "Handle and Transform Taxonomic Information"
#> [16] "Get inspired by real examples of how our packages can be used."
#> [17] "Or browse scientific publications that cited our packages."
#> [18] "Our suite of packages is comprised of contributions from staff engineers and the wider R\ncommunity via a transparent, constructive and open review process utilising GitHub's open\nsource infrastructure."
#> [19] "We combine academic peer reviews with production software code reviews to create a\ntransparent, collaborative & more efficient review process\n "
#> [20] "Based on best practices of software development and standards of R, its\napplications and user base."
#> [21] "Our diverse community of academics, data scientists and developers provide a\nplatform for shared learning, collaboration and reproducible science"
#> [22] "We welcome you to join us and help improve tools and practices available to\nresearchers while receiving greater visibility to your contributions. You can\ncontribute with your packages, resources or post questions so our members will help\nyou along your process."
#> [23] "Discover, learn and get involved in helping to shape the future of Data Science"
#> [24] "Join in our quarterly Community Calls with fellow developers and scientists - open\nto all"
#> [25] "Upcoming events including meetings at which our team members are speaking."
#> [26] "The latest developments from rOpenSci and the wider R community"
#> [27] "Release notes, updates and package related developements"
#> [28] "A digest of R package and software review news, use cases, blog posts, and events, curated every two weeks. Subscribe to get it in your inbox, or check the archive."
#> [29] "Happy rOpenSci users can be found at"
#> [30] "Except where otherwise noted, content on this site is licensed under the CC-BY license •\nPrivacy Policy"If needed, it’s possible to collapse the paragraphs into one bag of words:
paragraphs_scrap(link = "https://ropensci.org/", collapse = TRUE)
#> [1] " We help develop R packages for the sciences via community driven learning, review and\nmaintenance of contributed software in the R ecosystem Use our carefully vetted, staff- and community-contributed R software tools that lower barriers to working with local and remote scientific data sources. Combine our tools with the rich ecosystem of R packages. Workflow Tools for Your Code and Data Get Data from the Web Convert and Munge Data Document and Release Your Data Visualize Data Work with Databases From R Access, Manipulate, Convert Geospatial Data Interact with Web Resources Use Image & Audio Data Analyze Scientific Papers (and Text in General) Secure Your Data and Workflow Handle and Transform Taxonomic Information Get inspired by real examples of how our packages can be used. Or browse scientific publications that cited our packages. Our suite of packages is comprised of contributions from staff engineers and the wider R\ncommunity via a transparent, constructive and open review process utilising GitHub's open\nsource infrastructure. We combine academic peer reviews with production software code reviews to create a\ntransparent, collaborative & more efficient review process\n Based on best practices of software development and standards of R, its\napplications and user base. Our diverse community of academics, data scientists and developers provide a\nplatform for shared learning, collaboration and reproducible science We welcome you to join us and help improve tools and practices available to\nresearchers while receiving greater visibility to your contributions. You can\ncontribute with your packages, resources or post questions so our members will help\nyou along your process. Discover, learn and get involved in helping to shape the future of Data Science Join in our quarterly Community Calls with fellow developers and scientists - open\nto all Upcoming events including meetings at which our team members are speaking. The latest developments from rOpenSci and the wider R community Release notes, updates and package related developements A digest of R package and software review news, use cases, blog posts, and events, curated every two weeks. Subscribe to get it in your inbox, or check the archive. Happy rOpenSci users can be found at Except where otherwise noted, content on this site is licensed under the CC-BY license •\nPrivacy Policy"weblink_scrap()weblink_scrap() is used to srape the web links available within a web page. Useful in some cases, for example, getting a list of the available PDFs:
weblink_scrap(link = "https://www.worldbank.org/en/access-to-information/reports/",
contain = "PDF",
case_sensitive = FALSE)
#> Undefined Error: Error in open.connection(x, "rb"): Timeout was reached: [www.worldbank.org] Operation timed out after 10000 milliseconds with 0 out of 0 bytes received
#> [1] NAimages_scrap() and images_preview()images_preview() allows you to scrape the URLs of the images available within a web page so that you can choose which images extension (see below) you want to focus on.
Let’s say we want to list all the images from the official RStudio website:
images_preview(link = "https://rstudio.com/")
#> [1] "https://dc.ads.linkedin.com/collect/?pid=218281&fmt=gif"
#> [2] "https://www.facebook.com/tr?id=151855192184380&ev=PageView&noscript=1"
#> [3] "https://d33wubrfki0l68.cloudfront.net/08b39bfcd76ebaf8360ed9135a50a2348fe2ed83/75738/assets/img/logo-white.svg"
#> [4] "https://d33wubrfki0l68.cloudfront.net/8bd479afc1037554e6218c41015a8e047b6af0f2/d1330/assets/img/libertymutual-logo-regular.png"
#> [5] "https://d33wubrfki0l68.cloudfront.net/089844d0e19d6176a5c8ddff682b3bf47dbcb3dc/9ba69/assets/img/walmart-logo.png"
#> [6] "https://d33wubrfki0l68.cloudfront.net/a4ebff239e3de426fbb43c2e34159979f9214ce2/fabff/assets/img/janssen-logo-2.png"
#> [7] "https://d33wubrfki0l68.cloudfront.net/6fc5a4a8c3fa96eaf7c2dc829416c31d5dbdb514/0a559/assets/img/accenture-logo.png"
#> [8] "https://d33wubrfki0l68.cloudfront.net/d66c3b004735d83f205bc8a1c08dc39cc1ca5590/2b90b/assets/img/nasa-logo.png"
#> [9] "https://d33wubrfki0l68.cloudfront.net/521a038ed009b97bf73eb0a653b1cb7e66645231/8e3fd/assets/img/rstudio-icon.png"
#> [10] "https://d33wubrfki0l68.cloudfront.net/19dbfe44f79ee3249392a5effaa64e424785369e/91a7c/assets/img/connect-icon.png"
#> [11] "https://d33wubrfki0l68.cloudfront.net/edf453f69b61f156d1d303c9ebe42ba8dc05e58a/213d1/assets/img/icon-rspm.png"
#> [12] "https://d33wubrfki0l68.cloudfront.net/62bcc8535a06077094ca3c29c383e37ad7334311/a263f/assets/img/logo.svg"
#> [13] "https://d33wubrfki0l68.cloudfront.net/9249ca7ba197318b488c0b295b94357694647802/6d33b/assets/img/logo-lockup.svg"
#> [14] "https://d33wubrfki0l68.cloudfront.net/30ef84abbbcfbd7b025671ae74131762844e90a1/3392d/assets/img/bcorps-logo.svg"images_scrap() on the other hand download the images. It takes the following arguments:
link: The URL of the web page;
imgpath: The destination folder of your images. It defaults to getwd()
extn: the extension of the image: jpg, png, jpeg … among others;
askRobot: ask the robots.txt file if it’s permitted to scrape the web page.
In the following example we extract all the png images from RStudio :
# Suppose we're in a project which has a folder called my_images:
images_scrap(link = "https://rstudio.com/",
imgpath = here::here("my_images"),
extn = "png") # without the .The images will be downloaded into the folder here::here("myimages").
Please note that the ralger project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.