Navigating HTML

Web Scraping in R

Timo Grossenbacher

Instructor

HTML is like a tree

<html>
  <body>
    <div>
      <p>The first paragraph.</p>
    </div>
    <div>
      Not an actual paragraph, 
      but with a <a href="#">link</a>.
    </div>
    <p>A paragraph without an 
      enclosing div.</p>
  </body>
</html>

HTML tree

HTML is like a tree

<html>
  <body>
    <div>
      <p>The first paragraph.</p>
    </div>
    <div>
      Not an actual paragraph, 
      but with a <a href="#">link</a>.
    </div>
    <p>A paragraph without an 
      enclosing div.</p>
  </body>
</html>

HTML tree 2

HTML is like a tree

<html>
  <body>
    <div>
      <p>The first paragraph.</p>
    </div>
    <div>
      Not an actual paragraph, 
      but with a <a href="#">link</a>.
    </div>
    <p>A paragraph without an 
      enclosing div.</p>
  </body>
</html>

HTML tree 3

Navigating the tree with rvest

<html>
  <body>
    <div>
      <p>The first paragraph.</p>
    </div>
    <div>
      Not an actual paragraph, 
      but with a <a href="#">link</a>.
    </div>
    <p>A paragraph without an 
      enclosing div.</p>
  </body>
</html>

html <- read_html(html_document)

html_children(html)

{xml_nodeset (1)}
[1] <body>\n    <div>\n  < ...

html %>% html_children()

html %>% html_children() %>% html_text()

[1] "\n    \n      The first paragraph.\n   
\n    \n      Not an actual paragraph, \n      
but with a link.\n    \n    A paragraph ...

Navigating to nodes with selectors

<html>
  <body>
    <div>
      <p>The first paragraph.</p>
    </div>
    <div>
      Not an actual paragraph, 
      but with a <a href="#">link</a>.
    </div>
    <p>A paragraph without an 
      enclosing div.</p>
  </body>
</html>

html <- read_html(html_document)

html %>% html_element('body')

{xml_nodeset (1)}
[1] <body>\n    <div>\n  < ...

html %>% html_elements('div p')

{xml_nodeset (1)}
[1] <p>The first paragraph.</p>

Navigating to nodes with selectors

<html>
  <body>
    <div>
      <p>The first paragraph.</p>
    </div>
    <div>
      Not an actual paragraph, 
      but with a <a href="#">link</a>.
    </div>
    <p>A paragraph without an 
      enclosing div.</p>
  </body>
</html>

html %>% html_elements('p')

{xml_nodeset (2)}
[1] <p>The first paragraph.</p>
[2] <p>A paragraph without an enclosi...

html %>% html_elements('div') %>% 
    html_elements('p')

{xml_nodeset (1)}
[1] <p>The first paragraph.</p>

Extracting attributes

html %>%
    html_element('a') %>%
    html_attr('href')

[1] #

html %>%
    html_element('a') %>%
    html_attrs()

href 
 "#"

Let's do this!

Web Scraping in R