Scraping Amazon Product Data With Python: A Complete Guide

Amazon is packed with useful e-commerce data, such as product information, reviews, and prices. Extracting this data efficiently and putting it to use is imperative for any modern business. Whether you intend to monitor the performance of your products sold by third-party resellers or track your competitors, you need reliable web scraping services, like Amazon Scraper, to grasp this data for market analytics. 

Amazon scraping, however, has its peculiarities. In this step-by-step guide, we’ll go over every stage needed to create an Amazon web scraper. 

You can copy this user-agent and create a dictionary for the headers. 

The following example shows a dictionary with the user-agent and accept-language headers:

custom_headers = {
    'user-agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/108.0.0.0 Safari/537.36',
    'accept-language': 'en-GB,en;q=0.9',
}

You can send this dictionary to the optional parameter of the get method as follows:

response = requests.get(url, headers= custom_headers)

Executing the code with these changes may show the expected HTML with the product details.

Another note is that if you send as many headers as possible, you will not need Javascript rendering. If you need rendering, you will require tools like Playwright or Selenium. If the User-Agent and Accept-Language strings still bring you the 503 error, you can try to use the following headers:

custom_headers = {
    'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 13_1) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.1 Safari/605.1.15',
    'Accept-Language': 'da, en-gb, en',
    'Accept-Encoding': 'gzip, deflate, br',
    'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8',
    'Referer': 'https://www.google.com/'
}

It’s also a good idea to rotate different User-Agent strings and try your requests again to overcome the 503 error.

Scraping Amazon product data

When web scraping Amazon products, typically, you would work with two categories of pages — the category page and the product details page.

For example, open https://www.amazon.com/b?node=12097479011 or search for Over-Ear Headphones on Amazon. The page that shows the search results is the category page.

The category page displays the product title, product image, product rating, product price, and, most importantly, the product URLs page. If you want more details, such as product descriptions, you will get them only from the product details page.

Let's examine the structure of the product details page.

Open a product URL, such as https://www.amazon.com/Bose-QuietComfort-45-Bluetooth-Canceling-Headphones/dp/B098FKXT8L, in Chrome or any other modern browser, right-click the product title, and select Inspect. You will see that the HTML markup of the product title is highlighted.

You will see that it is a span tag with its id attribute set to "productTitle". 

Similarly, if you right-click the price and select Inspect, you will see the HTML markup of the price.

You can see that the dollar component of the price is in a span tag with the class "a-price-whole", and the cents component is in another span tag with the class set to "a-price-fraction". 

Similarly, you can locate the rating, image, and description.

Once you have this information, add the following lines to the code we have written so far:

1. Send a GET request with custom headers

from bs4 import BeautifulSoup

response = requests.get(url, headers=custom_headers)
soup = BeautifulSoup(response.text, 'lxml')

Beautiful Soup supports a unique way of selecting tags that utilize the find methods. Alternatively, Beautiful Soup also supports CSS selectors. You can use either of these to get the same results. In this guide, we will use CSS selectors, which are universal ways to select elements. CSS selectors work with almost all web scraping tools that can be used for web scraping Amazon product data.

We are now ready to use the Soup object to query for specific information.

2. Locate and scrape product name

The product name or the product title is located in a span element with its id productTitle. It's easy to select elements using the id that is unique.

See the following code for example:

title_element = soup.select_one('#productTitle')

We send the CSS selector to the select_one method, which returns an element instance.

We can extract information from the text using the text attribute.

title = title_element.text

Upon printing it, you will see that there are few white spaces. To fix that, add .strip() function call as follows:

title = title_element.text.strip()

3. Locate and scrape product rating

Scraping Amazon product ratings needs a little more work. 

First, let's create a selector for rating:

#acrPopover

Now, the following statement can select the element that contains the rating.

rating_element = soup.select_one('#acrPopover')

Note that the rating value is actually in the title attribute:

rating_text = rating_element.attrs.get('title')
print(rating_text)
# prints '4.6 out of 5 stars'

Lastly, we can use the replace method to get the number:

rating = rating_text.replace('out of 5 stars','')

4. Locate and scrape product price

The product price is located in two places — below the product title and also on the Buy Now box. 

We can use either of these tags to scrape Amazon product prices.

Let's create a CSS selector for the price:

span.a-offscreen

This CSS selector can be passed to the select_one method of BeautifulSoup as follows:

price_element = soup.select_one('span.a-offscreen')

You can now print the price 

print(price_element.text)

5. Locate and scrape product image

Let's scrape the default image. This image has the CSS selector as #landingImage. With this information, we can write the following lines of code to get the image URL from the src attribute:

image_element = soup.select_one('#landingImage')
image = image_element.attrs.get('src')

6. Locate and scrape product description

The next step in scraping Amazon product information is scraping the product description.

The methodology remains the same — create a CSS selector and use the select_one method.

The CSS selector for the description is as follows:

#productDescription

It means that we can extract the element as follows:

description_element = soup.select_one('#productDescription').text.strip()
print(description_element)

7. Locate and scrape reviews

Scraping Amazon product reviews necessitates a more extended explanation, so we covered it in a separate blog post. Please read it to learn more about locating and extracting data from Amazon review pages.

Handling product listing

So far, we have explored how to scrape product information.

However, to reach the product information, you will begin with product listing or category pages.

For example, https://www.amazon.com/b?node=12097479011 is the category page for over-ear headphones. 

If you examine this page, you will notice that all the products are contained in a div that has a special attribute [data-asin]. In that div, all the product links are in an h2 tag.

With this in mind, the CSS Selector would be as follows:

[data-asin] h2 a

We can read the href attribute of this selector and run a loop. However, note that the links will be relative. You would need to use the urljoin method to parse these links.

from urllib.parse import urljoin

def parse_listing(listing_url):
    global visited_urls
    response = requests.get(listing_url, headers=custom_headers)
    print(response.status_code)
    soup_search = BeautifulSoup(response.text, "lxml")
    link_elements = soup_search.select("[data-asin] h2 a")
    page_data = []

    for link in link_elements:
        full_url = urljoin(listing_url, link.attrs.get("href"))
        if full_url not in visited_urls:
            visited_urls.add(full_url)
            print(f"Scraping product from {full_url[:100]}", flush=True)
            product_info = get_product_info(full_url)
            if product_info:
                page_data.append(product_info)

Handling pagination

The link to the next page is in a link that contains the text Next. We can look for this link using the contains operator of CSS as follows:

    next_page_el = soup_search.select_one('a.s-pagination-next')
    if next_page_el:
        next_page_url = next_page_el.attrs.get('href')
        next_page_url = urljoin(listing_url, next_page_url)
        print(f'Scraping next page: {next_page_url}', flush=True)
        page_data += parse_listing(next_page_url)

    return page_data

8. Export scraped product data to a CSV file

The data we're scraping is being returned as a dictionary. This is intentional. We can create a list that contains all the scraped products:

def main():
    data = []
    search_url = "https://www.amazon.com/s?k=bose&rh=n%3A12097479011&ref=nb_sb_noss"
    data = parse_listing(search_url)

This page_data can then be used to create a Pandas DataFrame object:

    df = pd.DataFrame(data)
    df.to_csv("headphones.csv", index=False)

Reviewing final script 

Putting together everything, the following is the final script:

import requests
from bs4 import BeautifulSoup
from urllib.parse import urljoin
import pandas as pd

custom_headers = {
    'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 13_1) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.1 Safari/605.1.15',
    'Accept-Language': 'da, en-gb, en',
    'Accept-Encoding': 'gzip, deflate, br',
    'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8',
    'Referer': 'https://www.google.com/'
}

visited_urls = set()

def get_product_info(url):
    response = requests.get(url, headers=custom_headers)
    if response.status_code != 200:
        print(f"Error in getting webpage: {url}")
        return None

    soup = BeautifulSoup(response.text, "lxml")

    title_element = soup.select_one("#productTitle")
    title = title_element.text.strip() if title_element else None

    price_element = soup.select_one('span.a-offscreen')
    price = price_element.text if price_element else None

    rating_element = soup.select_one("#acrPopover")
    rating_text = rating_element.attrs.get("title") if rating_element else None
    rating = rating_text.replace("out of 5 stars", "") if rating_text else None

    image_element = soup.select_one("#landingImage")
    image = image_element.attrs.get("src") if image_element else None

    description_element = soup.select_one("#productDescription")
    description = description_element.text.strip() if description_element else None

    return {
        "title": title,
        "price": price,
        "rating": rating,
        "image": image,
        "description": description,
        "url": url
    }


def parse_listing(listing_url):
    global visited_urls
    response = requests.get(listing_url, headers=custom_headers)
    print(response.status_code)
    soup_search = BeautifulSoup(response.text, "lxml")
    link_elements = soup_search.select("[data-asin] h2 a")
    page_data = []

    for link in link_elements:
        full_url = urljoin(listing_url, link.attrs.get("href"))
        if full_url not in visited_urls:
            visited_urls.add(full_url)
            print(f"Scraping product from {full_url[:100]}", flush=True)
            product_info = get_product_info(full_url)
            if product_info:
                page_data.append(product_info)

    next_page_el = soup_search.select_one('a.s-pagination-next')
    if next_page_el:
        next_page_url = next_page_el.attrs.get('href')
        next_page_url = urljoin(listing_url, next_page_url)
        print(f'Scraping next page: {next_page_url}', flush=True)
        page_data += parse_listing(next_page_url)

    return page_data


def main():
    data = []
    search_url = "https://www.amazon.com/s?k=bose&rh=n%3A12097479011&ref=nb_sb_noss"
    data = parse_listing(search_url)
    df = pd.DataFrame(data)
    df.to_csv("headphones.csv", orient='records')


if __name__ == '__main__':
    main()

Best practices

Scraping Amazon without proxies or dedicated scraping tools is full of obstacles. Just like many other popular scraping targets, Amazon has rate-limiting in place, meaning it can block your IP address if you exceed the established limit. Apart from that, Amazon uses bot-detection algorithms that can check your HTTP headers for any suspicious details. Also, you should be ready to constantly adapt to the different page layouts and various HTML structures. 

Considering these factors, it’s recommended to follow some common practices to prevent getting detected and blocked by Amazon. Some of the most useful tips are: 

1. Use a real User-Agent. It’s important to make your User-Agent look as plausible as possible. Here’s the list of the most common user agents.

2. Set your fingerprint. Many websites use Transmission Control Protocol (TCP) and IP fingerprinting to detect bots. To avoid getting spotted, you need to make sure your fingerprint parameters are always consistent. 

3. Change the crawling pattern. To develop a successful crawling pattern, you should think about how a regular user would behave while exploring a page and add clicks, scrolls, and mouse movements accordingly.

And this is only a small portion of the requirements you should keep in mind when scraping Amazon.

An easier solution to extract Amazon data 

Alternatively, you can turn to a ready-made scraping solution designed specifically for scraping Amazon - Amazon Scraper API. With this scraper, you can:

• Scrape and parse various Amazon page types, including Search, Product, Offer listing, Questions & Answers, Reviews, Best Sellers, and Sellers.

• Target localized product data in 195 locations worldwide;

• Retrieve accurate parsed results in JSON format without installing any other library;

• Enjoy multiple handy features, such as bulk scraping and automated jobs.

Let's look at Amazon Scraper API in action.

Scraping products from search results

You can search and extract the products from Amazon with this straightforward code example:

import requests
from pprint import pprint

# Structure payload.
payload = {
    'source': 'amazon_search',
    'query': 'bose',  # Search for "bose"
    'start_page': 1,
    'pages': 10,
    'parse': True,
    'context': [
        {'key': 'category_id', 'value': 12097479011}  # category id for headphones
    ],
}

# Get response
response = requests.request(
    'POST',
    'https://realtime.oxylabs.io/v1/queries',
    auth=('USERNAME', 'PASSWORD'),
    json=payload,
)

# Print prettified response to stdout.
pprint(response.json())

Notice how it requests 10 pages beginning with page 1. Also, we limit the search to category ID 12097479011, which is Amazon's category ID for headphones. You’ll get the data returned in JSON format: