search docs

Web scraping techniques

An introduction to the methods you can use to extract data from websites. Analyze web pages for hidden elements to find the most effective approach.

This article provides a quick summary of ways websites structure and send their information. Knowing these techniques will help you extract data quicker and more efficiently.

You can find more in-depth discussion and code examples for all the techniques in this article.

CSS selectors

The first method you will use when trying web scraping is most likely CSS selectors. They allow you to select your desired elements by type, class, ID or attributes.

To see which element contains the information you need (and its details), open your browser's Developer Tools. Right-click the content, then select Inspect on Chrome and Inspect element on Mozilla.

CSS selectors

If you are only looking to scrape a couple of elements from a page, this method is sufficient. For more elaborate extraction use cases, however, there are other, more effective, methods. microdata

Schemas provide a way to mark up web pages so major search engines like Google, Bing and Yahoo can understand them.

Pages with markup still use HTML. The only difference is that they add machine-readable code markers into the HTML documents. This helps cut down on ambiguity and allows search engines to return more accurate results.

If a site uses microdata, you will find it in its <head> element using your browser's developer tools. It will be in a hidden element similar to the one below.

<div itemscope itemtype="">
  <span>Director: Mati Diop</span>
  <a href="../movies/atlantics-theatrical-trailer.html">Trailer</a>

Check out this tutorial to learn how to scrape pages using microdata.


Similar to microdata, some sites use JSON for Linking Data (JSON-LD). Based on the JSON format, JSON-LD helps structure a web page's content in a way that's easy for humans and computers to read. This improves a site's SEO.

To see if a website uses JSON-LD, check its <head> element using your browser's developer tools. You will find the JSON-LD data in a <script> element similar to the one below.

JSON-LD data

Then, you can use the code below in your scraper to extract the data.

const pageFunction = (context) => {
    // other code here
    const jsonLD = $('script[type="application/ld+json"]');
    return JSON.parse(jsonLD.html());

Internal JavaScript variables

You can find internal JavaScript variables in a <script> tag in a web page's <head> element. These variables contain custom data that is set when the page loads.

To use internal variables, first load the page using headless Chrome (usually with Puppeteer). Once it is loaded, you can return the variable that contains it using return {variable_name} (within the page context).

For example, if you are searching for Kaffeine café on Yelp, search for the café's name in the page's HTML using developer tools. The data you would receive would look something like the example below.

    "city": "London",
    "category_aliases": "cafe",
    "biz_id": "iUockw0CUssKZLyoGJYEXA",
    "biz_name": "Kaffeine London",
    "latitude": 51.5074,
    "longitude": 0.1278,

Check out this tutorial for more information and code examples for scraping with internal JavaScript variables.


XMLHttpRequests (XHRs) allow you to scrape websites that load their data dynamically. When you perform the required action (e.g. click a Load more button), the page makes an API call to the XHR, which returns the data that will be displayed.

Using XHRs, you can access an entire API's content without even looking at the page. Just find the XHR that retrieves the data you want and use it to retrieve the data in a tidy structured format.

You can check the XHRs a web page sends under the Network tab, in the XHR section in your browser's developer tools.

Network tab and XHR section

Check out this tutorial for more information and code examples for scraping with XHRs.