Future of Web Scraping

The Internet is large, complex and ever-evolving. Nearly 90% of all the data in the world has been generated over the last two years. In this vast ocean of data, how does one get to the relevant piece of information? This is where web scraping takes over.

Web scrapers attach themselves, like a leech, to this beast and ride the waves by extracting information form websites at will. Granted “scraping” doesn’t have a lot of positive connotations, yet it happens to be the only way to access data or content from a web site without RSS or an open API.

Future of Web Scraping

Web scraping faces testing times ahead. We outline why there may be some serious challenges to its future.

With rise in data, redundancies in web scraping are rising. No more is web scraping a domain of the coders; in fact, companies now offer customized scraping tools to clients which they can use to get the data they want. The outcome of everyone equipped to crawl, scrape, and extract, is unnecessary waste of precious man-power. Collaborative scraping could well heal this hurt. Here, where one web crawler does a broad scraping, the others scrape data off an API. An extension of the problem is that text retrieval attracts more attention than multimedia; and with websites becoming more complex, this enforces limited scraping capacity.

Easily, the biggest challenge to web scraping technology is Privacy concerns. With data freely available (most of it voluntary, much of it involuntary), the call for stricter legislation rings loudest. Unintended users can easily target a company and take advantage of the business using web scraping. The disdain with which “do not scrape” policies are treated and terms of usage violated, tells us that even legal restrictions are not enough. This begs to ask an age-old question: is scraping legal?

Is Crawling Legal? from PromptCloud

The flipside to this argument is that if technological barriers replace legal clauses, then web scraping will see a steady, and sure, decline. This is a distinct possibility since the only way scraping activity thrives is on the grid, and if the very means are taken away and programs no longer have access to website information, then web scraping by itself will be wiped out.

Building the Future

On the same thought is the growing trend of accepting “open data”. The open data policy, while long mused hasn’t been used at the scale it should be. The old way was to believe that closed data is the edge over competitors. But that mindset is changing. Increasingly, websites are beginning to offer APIs and embracing open data. But what’s the advantage of doing so?

Selling APIs not only brings in the money, but also is useful in driving back traffic to the sites! APIs are also a more controlled, cleaner way of turning sites into services. Steadily many successful sites like Twitter, LinkedIn etc. are offering access to their APIs with paid services and actively blocking scraper and bots.

Yet, beyond these obvious challenges, there’s a glimmer of hope for web scraping. And this is based on a singular factor: the growing need for data!

With Internet & web technology spreading, massive amounts of data will be accessible on the web. Particularly with increased adoption of mobile internet. According to one report, by 2020, the number of mobile internet users will hit 3.8 billion, or around half of the world’s population!

Since ‘big data’ can be both, structured & unstructured; web scraping tools will only get sharper and incisive. There is fierce competition between those who provide web scraping solutions. With the rise of open source languages like Python, R & Ruby, Customized scraping tools will only flourish bringing in a new wave of data collection and aggregation methods.

Source: https://www.promptcloud.com/blog/Future-of-Web-Scraping

3 Reasons to Up Your Web Scraping Game

If you aren’t using a machine-learning-driven intelligent Web scraping solution yet, here are three reasons why you might want to abandon that entry-level Web-scraping software or cut your high-cost script-writing approach.

    You need to keep an eye on a large number of web sources that get updated frequently.
    Understanding what’s changed is at least as critical as the data itself.
    You don’t want maintenance and scheduling to drag you down.

Here’s what an intelligent Web-scraping solution can deliver – and why:

1. Better data monitoring of an ever-shifting Web

If you need to keep a watch over hundreds, thousands or even tens of thousands of sites, an intelligent Web scraper is a must, because:

    It can scale – easily adding new websites, coordinating extraction routines, and automating the normalization of data across different websites.

    It can navigate and extract data from websites efficiently. Script-based approaches typically only can view a Web page in isolation, making it difficult to optimize navigation across unique pages of a targeted site. More intelligent approaches can be trained to bypass unnecessary links and leave a lighter footprint on the sites you need to access. And, they can monitor millions of precise Web data points quickly. This means you can monitor more pages on more sites with more frequent updates.

2. Critical alerts to Web data changes

A key sales executive suddenly drops off of the management page of your main competitor. That can mean big shakeup in the entire organization, which your sales team can jump on.

An intelligent Web scraper can alert you to this personnel shift because it can be set to monitor for just the changes; less powerful technologies or script-based approaches can’t. Whether you’re tracking price shifts, people moves, or product changes (or more) intelligent Web scraping delivers more profound insights.

3. Maintenance may become your biggest nightmare

You’ve purchased an entry-level tool and built out scrapers for a few hundred sites.  At first, everything seems fine. But, within weeks you begin to notice that your data is incomplete and not being updated as you’d expected. Why did your data deliveries disappear?

Reality is that these low-cost tools are simply not designed for mission-critical business applications – on the surface they look helpful and easy to use, but underneath the surface they are script-based and highly dependent upon the HTML of a website. But websites change, and entry-level web scraping tools are simply not engineered to adapt to those changes.

And, most of these tools are simply not designed for enterprise use. They have limited reporting, if any, so the only way to know whether they’re successfully completing their tasks is by finding gaps in the data – often when it’s too late.

An intelligent web scraping approach doesn’t rely upon the HTML of a web page. It uses machine learning algorithms which view the web the same way a user might. A typical reader doesn’t get confused when a font or color is changed on a website, and neither do these algorithms. But simple approaches to web scraping are highly dependent on the specific HTML to help it understand the content of a page. So, when websites have design changes (on average once every 18 months), the software fails.

While entry-level web scraping software can be an easy solution for simple, one-time web scraping projects, the scripts they generate are fragile and the resources required for tracking and maintenance can become overwhelming when you need to regularly extract data from multiple sites.

Case in point: Shopzilla assimilates data five times faster than outsourced Web scrapers

To demonstrate the power of intelligent Web scraping, here’s a real-life example from Shopzilla.  Shopzilla manages a premier portfolio of online shopping brands in the United States and Europe, connecting more than 40 million shoppers each month with millions of products from retailers worldwide. With the explosive growth of retail data on the Web, Shopzilla’s outsourced, custom-built approach, based on scripting, could not add the product lines of new retailers to its site in a timely fashion. It was taking up to two weeks to write the scripts needed to make a single site accessible.

By deploying Connotate’s intelligent web scraping platform on site, Shopzilla gained the ability to harness Web data’s rapid growth and keep up to date. Today, new sources are added in days, not weeks.  The platform continually monitors Web content from thousands of sites, delivering high volumes of data every day in a structured format. The result: 500 percent more data from new retailers. An added bonus: the company has reduced IT maintenance costs and its dependence on outsourced development timetables. Case in point: Deep competitor intelligence in two languages

A global manufacturer needed to monitor competitors’ technology improvements in a field where market leadership hinges on an ability to quickly leverage these advances. That meant accessing scholarly journals and niche sites in multiple languages. Using the Connotate solution, it was able to access highly-targeted, keyword-driven university and industry research journals and blogs in German and English that are hard to reach because they do not support RSS feeds. Our solution also incorporated semantic analysis to tag and categorize data and help identify new technologies and products not currently in the keyword list. The firm enhanced its competitive edge with the up-to-the-minute, precise data it needed.

Is your Web scraping intelligent enough?

See what intelligent agents through an automated Web data extraction and monitoring solution can bring to your business. Contact us and speak with one of experts.

Source:http://www.connotate.com/3-reasons-web-scraping-game-6579#.VGMjH2f4EuQ

Web Scraping Popularity Soars

The world is stirred because of the ever-growing web scraping success in almost all of its services. Success stories pertaining to the benefits of online data collection in business, research, politics, health, and almost all aspects of human life are endless. With this popularity surge, it has become a hot issue and many are questioning its legality and reliability.

Looking back, this simple harvesting of pertinent data from competitors and the global market in general like anything else started as a non-threatening and advanced form of web research. Eventually, when the benefits begin to manifest and the system improves, many are lured into it that it has become one of the strongest and fastest growing business in the world.

Simple Beginnings

As naturally as a law of life that great things come from small beginnings, data mining was conceived as a process in gaining information, mostly in research. This act of collecting information through the internet was never imagined to be what it has become nowadays.

Source:http://www.loginworks.com/blogs/web-scraping-blogs/web-scraping-popularity-soars/

Scraping webdata from a website that loads data in a streaming fashion

I'm trying to scrape some data off of the FEC.gov website using python for a project of mine. Normally I use python

mechanize and beautifulsoup to do the scraping.

I've been able to figure out most of the issues but can't seem to get around a problem. It seems like the data is

streamed into the table and mechanize.Browser() just stops listening.

So here's the issue: If you visit http://query.nictusa.com/cgi-bin/can_ind/2011_P80003338/1/A ... you get the first 500

contributors whose last name starts with A and have given money to candidate P80003338 ... however, if you use

browser.open() at that url all you get is the first ~5 rows.

I'm guessing its because mechanize isn't letting the page fully load before the .read() is executed. I tried putting a

time.sleep(10) between the .open() and .read() but that didn't make much difference.

And I checked, there's no javascript or AJAX in the website (or at least none are visible when you use the 'view-

source'). SO I don't think its a javascript issue.

Any thoughts or suggestions? I could use selenium or something similar but that's something that I'm trying to avoid.

-Will

2 Answers

Why not use an html parser like lxml with xpath expressions.

I tried

>>> import lxml.html as lh
>>> data = lh.parse('http://query.nictusa.com/cgi-bin/can_ind/2011_P80003338/1/A')
>>> name = data.xpath('/html/body/table[2]/tr[5]/td[1]/a/text()')
>>> name
[' AABY, TRYGVE']
>>> name = data.xpath('//table[2]/*/td[1]/a/text()')
>>> len(name)
500
>>> name[499]
' AHMED, ASHFAQ'
>>>



Similarly, you can create xpath expression of your choice to work with.


Source: http://stackoverflow.com/questions/9435512/scraping-webdata-from-a-website-that-loads-data-in-a-streaming-

fashion

How can I circumvent page view limits when scraping web data using Python?

I am using Python to scrape US postal code population data from http:/www.city-data.com, through this directory: http://www.city-data.com/zipDir.html. The specific pages I am trying to scrape are individual postal code pages with URLs like this: http://www.city-data.com/zips/01001.html. All of the individual zip code pages I need to access have this same URL Format, so my script simply does the following for postal_code in range:

    Creates URL given postal code
    Tries to get response from URL
    If (2), Check the HTTP of that URL
    If HTTP is 200, retrieves the HTML and scrapes the data into a list
    If HTTP is not 200, pass and count error (not a valid postal code/URL)
    If no response from URL because of error, pass that postal code and count error
    At end of script, print counter variables and timestamp

The problem is that I run the script and it works fine for ~500 postal codes, then suddenly stops working and returns repeated timeout errors. My suspicion is that the site's server is limiting the page views coming from my IP address, preventing me from completing the amount of scraping that I need to do (all 100,000 potential postal codes).

My question is as follows: Is there a way to confuse the site's server, for example using a proxy of some kind, so that it will not limit my page views and I can scrape all of the data I need?

Thanks for the help! Here is the code:

##POSTAL CODE POPULATION SCRAPER##

import requests

import re

import datetime

def zip_population_scrape():

    """
    This script will scrape population data for postal codes in range
    from city-data.com.
    """
    postal_code_data = [['zip','population']] #list for storing scraped data

    #Counters for keeping track:
    total_scraped = 0
    total_invalid = 0
    errors = 0


    for postal_code in range(1001,5000):

        #This if statement is necessary because the postal code can't start
        #with 0 in order for the for statement to interate successfully
        if postal_code <10000:
            postal_code_string = str(0)+str(postal_code)
        else:
            postal_code_string = str(postal_code)

        #all postal code URLs have the same format on this site
        url = 'http://www.city-data.com/zips/' + postal_code_string + '.html'

        #try to get current URL
        try:
            response = requests.get(url, timeout = 5)
            http = response.status_code

            #print current for logging purposes
            print url +" - HTTP:  " + str(http)

            #if valid webpage:
            if http == 200:

                #save html as text
                html = response.text

                #extra print statement for status updates
                print "HTML ready"

                #try to find two substrings in HTML text
                #add the substring in between them to list w/ postal code
                try:           

                    found = re.search('population in 2011:</b> (.*)<br>', html).group(1)

                    #add to # scraped counter
                    total_scraped +=1

                    postal_code_data.append([postal_code_string,found])

                    #print statement for logging
                    print postal_code_string + ": " + str(found) + ". Data scrape successful. " + str(total_scraped) + " total zips scraped."
                #if substrings not found, try searching for others
                #and doing the same as above   
                except AttributeError:
                    found = re.search('population in 2010:</b> (.*)<br>', html).group(1)

                    total_scraped +=1

                    postal_code_data.append([postal_code_string,found])
                    print postal_code_string + ": " + str(found) + ". Data scrape successful. " + str(total_scraped) + " total zips scraped."

            #if http =404, zip is not valid. Add to counter and print log        
            elif http == 404:
                total_invalid +=1

                print postal_code_string + ": Not a valid zip code. " + str(total_invalid) + " total invalid zips."

            #other http codes: add to error counter and print log
            else:
                errors +=1

                print postal_code_string + ": HTTP Code Error. " + str(errors) + " total errors."

        #if get url fails by connnection error, add to error count & pass
        except requests.exceptions.ConnectionError:
            errors +=1
            print postal_code_string + ": Connection Error. " + str(errors) + " total errors."
            pass

        #if get url fails by timeout error, add to error count & pass
        except requests.exceptions.Timeout:
            errors +=1
            print postal_code_string + ": Timeout Error. " + str(errors) + " total errors."
            pass


    #print final log/counter data, along with timestamp finished
    now= datetime.datetime.now()
    print now.strftime("%Y-%m-%d %H:%M")
    print str(total_scraped) + " total zips scraped."
    print str(total_invalid) + " total unavailable zips."
    print str(errors) + " total errors."



Source: http://stackoverflow.com/questions/25452798/how-can-i-circumvent-page-view-limits-when-scraping-web-data-using-python

Web data scraping (online news comments) with Scrapy (Python)

Since you seem like the try-first ask-question later type (that's a very good thing), I won't give you an answer, but a

(very detailed) guide on how to find the answer.

The thing is, unless you are a yahoo developer, you probably don't have access to the source code you're trying to

scrape. That is to say, you don't know exactly how the site is built and how your requests to it as a user are being

processed on the server-side. You can, however, investigate the client-side and try to emulate it. I like using Chrome

Developer Tools for this, but you can use others such as FF firebug.

So first off we need to figure out what's going on. So the way it works, is you click on the 'show comments' it loads

the first ten, then you need to keep clicking for the next ten comments each time. Notice, however, that all this

clicking isn't taking you to a different link, but lively fetches the comments, which is a very neat UI but for our

case requires a bit more work. I can tell two things right away:

    They're using javascript to load the comments (because I'm staying on the same page).
    They load them dynamically with AJAX calls each time you click (meaning instead of loading the comments with the

page and just showing them to you, with each click it does another request to the database).

Now let's right-click and inspect element on that button. It's actually just a simple span with text:

<span>View Comments (2077)</span>

By looking at that we still don't know how that's generated or what it does when clicked. Fine. Now, keeping the

devtools window open, let's click on it. This opened up the first ten. But in fact, a request was being made for us to

fetch them. A request that chrome devtools recorded. We look in the network tab of the devtools and see a lot of

confusing data. Wait, here's one that makes sense:

http://news.yahoo.com/_xhr/contentcomments/get_comments/?content_id=42f7f6e0-7bae-33d3-aa1d-

3dfc7fb5cdfc&_device=full&count=10&sortBy=highestRated&isNext=true&offset=20&pageNumber=2&_media.modules.content_commen

ts.switches._enable_view_others=1&_media.modules.content_comments.switches._enable_mutecommenter=1&enable_collapsed_com

ment=1

See? _xhr and then get_comments. That makes a lot of sense. Going to that link in the browser gave me a JSON object

(looks like a python dictionary) containing all the ten comments which that request fetched. Now that's the request you

need to emulate, because that's the one that gives you what you want. First let's translate this to some normal reqest

that a human can read:

go to this url: http://news.yahoo.com/_xhr/contentcomments/get_comments/
include these parameters: {'_device': 'full',
          '_media.modules.content_comments.switches._enable_mutecommenter': '1',
          '_media.modules.content_comments.switches._enable_view_others': '1',
          'content_id': '42f7f6e0-7bae-33d3-aa1d-3dfc7fb5cdfc',
          'count': '10',
          'enable_collapsed_comment': '1',
          'isNext': 'true',
          'offset': '20',
          'pageNumber': '2',
          'sortBy': 'highestRated'}

Now it's just a matter of trial-and-error. However, a few things to note here:

    Obviously the count is what decides how many comments you're getting. I tried changing it to 100 to see what

happens and got a bad request. And it was nice enough to tell me why - "Offset should be multiple of total rows". So

now we understand how to use offset

    The content_id is probably something that identifies the article you are reading. Meaning you need to fetch that

from the original page somehow. Try digging around a little, you'll find it.

    Also, you obviously don't want to fetch 10 comments at a time, so it's probably a good idea to find a way to fetch

the number of total comments somehow (either find out how the page gets it, or just fetch it from within the article

itself)

    Using the devtools you have access to all client-side scripts. So by digging you can find that that link to

/get_comments/ is kept within a javascript object named YUI. You can then try to understand how it is making the

request, and try to emulate that (though you can probably figure it out yourself)

    You might need to overcome some security measures. For example, you might need a session-key from the original

article before you can access the comments. This is used to prevent direct access to some parts of the sites. I won't

trouble you with the details, because it doesn't seem like a problem in this case, but you do need to be aware of it in

case it shows up.

    Finally, you'll have to parse the JSON object (python has excellent built-in tools for that) and then parse the

html comments you are getting (for which you might want to check out BeautifulSoup).

As you can see, this will require some work, but despite all I've written, it's not an extremely complicated task

either.

So don't panic.

It's just a matter of digging and digging until you find gold (also, having some basic WEB knowledge doesn't hurt).

Then, if you face a roadblock and really can't go any further, come back here to SO, and ask again. Someone will help

you.


Source: http://stackoverflow.com/questions/20218855/web-data-scraping-online-news-comments-with-scrapy-python

A good web data extraction/screen scraper program?

I need to capture product data from a site on a regular basis and wondered if any one knows of a good software program? I've trialed Mozenda but its a monthly subscription and pricey in the long term. Obviously something thats free would be best but I don't mind paying either. Just need a decent program thats reliable and doesn't require much programming knowledge.

You can try ScraperWiki.com if you know python.

I've experimented with Screen-Scraper and found it easy to use. The application comes in multiple versions: basic (which is free), professional, and enterprise. Also, multiple platforms are supported.

Hire a programmer to do it so that there is only a one off cost. I often see similar projects on freelancing websites like Elance and oDesk.

I really like iMacros. You can give it a test drive to see if it meets your needs with the totally free Firefox extension (there's also IE versions), but there are also more full featured application and "server" versions that have more features and ability to do thing in an unattended manner.

Here are some other alternatives to consider:

    License the data from the provider. Call em up and ask 'em.

    Use Amazon Mechanical Turk to get humans to copy and paste and format it for ya. They are cheap.

    For automation, it depends on how complicated the HTML is and how often it changes. You could use Excel's Web Data Import if it's really simple.


You can use irobot from IRobotSoft, which is totally free, and provides more functionalityies than other paid software. Watch demos here http://irobotsoft.com/help/ for how simple it is.

Questions on their forum were answered very quickly.


Source: http://stackoverflow.com/questions/2334164/a-good-web-data-extraction-screen-scraper-program