Scraping R-bloggers with Python – Part 2

In my previous post I showed how to write a small simple python script to download the pages of R-bloggers.com. If you followed that post and ran the script, you should have a folder on your hard drive with 2409 .html files labeled post1.html , post2.html and so forth. The next step is to write a small script that extract the information we want from each page, and store that information in a .csv file that is easily read by R. In this post I will show how to extract the post title, author name and date of a given post and store it in a .csv file with a unique id.

To do this open a document in your favorite python editor (I like to use aquamacs) and name it: extraction.py. As in the previous post we start by importing the modules that we will use for the extraction:

from BeautifulSoup import BeautifulSoup

import os
import re

As in the previous post we will be using the BeautifulSoup module to extract the relevant information from the pages. The os module is used to get a list of file from the directory where we have saved the .html files, and finally the re module allows us to use regular expressions to format the titles that include a comma value or a newline value (\n). We need to remove these as they would mess up the formatting of the .csv file.

After having read in the modules, we need to get a list of files that we can iterate over. First we need to specify the path were the files are saved, and then we use the os module to get all the filenames in the specified directory:

path = "/Users/thomasjensen/Documents/RBloggersScrape/download"

listing = os.listdir(path)

It might be that there are other files in the given directory, hence we apply a filter, in shape of a list comprehension, to weed out any file names that do not match our naming scheme:

listing = [name for name in listing if re.search(r"post\d+\.html",name) != None]

Notice that a regular expression was used to determine whether a given name in the list matched our naming scheme. For more on regular expressions have a look at this site.

The final steps in preparing our extraction is to change the working directory to where we have our .html files, and create an empty dictionary:

os.chdir(path)
data = {}

Dictionaries are one of the great features of Python. Essentially a dictionary is a mapping of a key to a specific value, however the fact that dictionaries can be nested within each other, allows us to create data structures similar to R’s data frames.

Now we are ready to begin extracting information from our downloaded pages. Much as in the previous post, we will loop over all the file names, read each file into Python and create a BeautifulSoup object from the file:

for page in listing:
    site = open(page,"rb")
    soup = BeautifulSoup(site)

In order to store the values we extract from a given page, we update the dictionary with a unique key for the page. Since our naming scheme made sure that each file had a unique name, we simply remove the .html part from the page name, and use that as our key:

key = re.sub(".html","",page)

data.update({key:{}})

This will create a mapping between our key and an empty dictionary, nested within the data dictionary. Once this is done we can start extract information and store it in our newly created nested dictionary. The content we want is located in the main column, which has the id tag “leftcontent” in the html code. To get at this we use find() function on soup object created above:

content = soup.find("div", id = "leftcontent")

The first “h1” tag in our content object contains the title, so again we will use the find() function on the content object, to find the first “h1” tag:

title = content.findNext("h1").text

To get the text within the “h1” tag the .text had been added to our search with in the content object.

To find the author name, we are lucky that there is a class of “div” tags called “meta” which contain a link with the author name in it. To get the author name we simply find the meta div class and search for a link. Then we pull out the text of the link tag:

author = content.find("div",{"class":"meta"}).findNext("a").text

Getting the date is a simple matter as it is nested within div tag with the class “date”:

date = content.find("div",{"class":"date"}).text

Once we have the three variables we put them in dictionaries that are nested within the nested dictionary we created with the key:

data[key]["title"] = title
data[key]["author"] = author
data[key]["date"] = date

Once we have run the loop and gone through all posts, we need to write them in the right format to a .csv file. To begin with we open a .csv file names output:

output = open("/Users/thomasjensen/Documents/RBloggersScrape/output.csv","wb")

then we create a header that contain the variable names and write it to the output.csv file as the first row:

variables = unicode(",".join(["id","date","author","title"]))
header = variables + "\n"
output.write(header.encode("utf8"))

Next we pull out all the unique keys from our dictionary that represent individual posts:

keys = data.keys()

Now it is a simple matter of looping through all the keys, pull out the information associated with each key, and write that information to the output.csv file:

for key in keys:
    print key
    id = key
    date = re.sub(",","",data[key]["date"])
    author = data[key]["author"]
    title = re.sub(",","",data[key]["title"])
    title = re.sub("\\n","",title)
    linelist = [id,date,author,title]
    linestring = unicode(",".join(linelist))
    linestring = linestring + "\n"
    output.write(linestring.encode("utf-8"))

Notice that we first create four variables that contain the id, date, author and title information. With regards to the title we use two regular expressions to remove any commas and “\n” from the title, as these would create new columns or new line breaks in the output.csv file. Finally we put the variables together in a list, and turn the list into a string with the list items separated by a comma. Then a linebreak is added to the end of the string, and the string is written to the output.csv file. As a last step we close the file connection:

output.close()

And that is it. If you followed the steps you should now have a csv file in your directory with 2409 rows, and four variables – ready to be read into R. Stay tuned for the next post which will show how we can use this data to see how R-bloggers has developed since 2005. The full extraction script is shown below:

from BeautifulSoup import BeautifulSoup

import os
import re

 path = "/Users/thomasjensen/Documents/RBloggersScrape/download"
 listing = os.listdir(path)

listing = [name for name in listing if re.search(r"post\d+\.html",name) != None]
 os.chdir(path)
 data = {}
 for page in listing:
site = open(page,"rb")
soup = BeautifulSoup(site)
key = re.sub(".html","",page)
print key
data.update({key:{}})
 content = soup.find("div", id = "leftcontent")
title = content.findNext("h1").text
author = content.find("div",{"class":"meta"}).findNext("a").text
date = content.find("div",{"class":"date"}).text
data[key]["title"] = title
data[key]["author"] = author
data[key]["date"] = date

 output = open("/Users/thomasjensen/Documents/RBloggersScrape/output.csv","wb")

 keys = data.keys()
 variables = unicode(",".join(["id","date","author","title"]))
 header = variables + "\n"
 output.write(header.encode("utf8"))
 for key in keys:
print key
id = key
date = re.sub(",","",data[key]["date"])
author = data[key]["author"]
title = re.sub(",","",data[key]["title"])
title = re.sub("\\n","",title)
linelist = [id,date,author,title]
linestring = unicode(",".join(linelist))
linestring = linestring + "\n"
output.write(linestring.encode("utf-8"))
 output.close()

Source:http://www.r-bloggers.com/scraping-r-bloggers-with-python-part-2/

Data Mining and Frequent Datasets

I've been doing some work for my exams in a few days and I'm going through some past papers but unfortunately there are no corresponding answers. I've answered the question and I was wondering if someone could tell me if I am correct.

My question is

    (c) A transactional dataset, T, is given below:
    t1: Milk, Chicken, Beer
    t2: Chicken, Cheese
    t3: Cheese, Boots
    t4: Cheese, Chicken, Beer,
    t5: Chicken, Beer, Clothes, Cheese, Milk
    t6: Clothes, Beer, Milk
    t7: Beer, Milk, Clothes

    Assume that minimum support is 0.5 (minsup = 0.5).

    (i) Find all frequent itemsets.

Here is how I worked it out:

    Item : Amount
    Milk : 4
    Chicken : 4
    Beer : 5
    Cheese : 4
    Boots : 1
    Clothes : 3

Now because the minsup is 0.5 you eliminate boots and clothes and make a combo of the remaining giving:

    {items} : Amount
    {Milk, Chicken} : 2
    {Milk, Beer} : 4
    {Milk, Cheese} : 1
    {Chicken, Beer} : 3
    {Chicken, Cheese} : 3
    {Beer, Cheese} : 2

Which leaves milk and beer as the only frequent item set then as it is the only one above the minsup?

data mining

Nanor

3 Answers

There are two ways to solve the problem:

    using Apriori algorithm
    Using FP counting

Assuming that you are using Apriori, the answer you got is correct.

The algorithm is simple:

First you count frequent 1-item sets and exclude the item-sets below minimum support.

Then count frequent 2-item sets by combining frequent items from previous iteration and exclude the item-sets below support threshold.

The algorithm can go on until no item-sets are greater than threshold.

In the problem given to you, you only get 1 set of 2 items greater than threshold so you can't move further.

There is a solved example of further steps on Wikipedia here.

You can refer "Data Mining Concepts and Techniques" by Han and Kamber for more examples.

141

There is more than two algorithms to solve this problem. I will just mention a few of them: Apriori, FPGrowth, Eclat, HMine, DCI, Relim, AIM, etc. –  Phil Mar 5 '13 at 7:18

OK to start, you must first understand, data mining (sometimes called data or knowledge discovery) is the process of analyzing data from different perspectives and summarizing it into useful information - information that can be used to increase revenue, cuts costs, or both. Data mining software is one of a number of analytical tools for analyzing data. It allows users to analyze data from many different dimensions or angles, categorize it, and summarize the relationships identified. Technically, data mining is the process of finding correlations or patterns among dozens of fields in large relational databases.

Now, the amount of raw data stored in corporate databases is exploding. From trillions of point-of-sale transactions and credit card purchases to pixel-by-pixel images of galaxies, databases are now measured in gigabytes and terabytes. (One terabyte = one trillion bytes. A terabyte is equivalent to about 2 million books!) For instance, every day, Wal-Mart uploads 20 million point-of-sale transactions to an A&T massively parallel system with 483 processors running a centralized database.

Raw data by itself, however, does not provide much information. In today's fiercely competitive business environment, companies need to rapidly turn these terabytes of raw data into significant insights into their customers and markets to guide their marketing, investment, and management strategies.

Now you must understand that association rule mining is an important model in data mining. Its mining algorithms discover all item associations (or rules) in the data that satisfy the user-specified minimum support (minsup) and minimum confidence (minconf) constraints. Minsup controls the minimum number of data cases that a rule must cover. Minconf controls the predictive strength of the rule.

Since only one minsup is used for the whole database, the model implicitly assumes that all items in the data are of the same nature and/or have similar frequencies in the data. This is, however, seldom the case in real- life applications. In many applications, some items appear very frequently in the data, while others rarely appear. If minsup is set too high, those rules that involve rare items will not be found. To find rules that involve both frequent and rare items, minsup has to be set very low.

This may cause combinatorial explosion because those frequent items will be associated with one another in all possible ways. This dilemma is called the rare item problem. This paper proposes a novel technique to solve this problem. The technique allows the user to specify multiple minimum supports to reflect the natures of the items and their varied frequencies in the database. In rule mining, different rules may need to satisfy different minimum supports depending on what items are in the rules.

Given a set of transactions T (the database), the problem of mining association rules is to discover all association rules that have support and confidence greater than the user-specified minimum support (called minsup) and minimum confidence (called minconf).

I hope that once you understand the very basics of data mining that the answer to this question shall become apparent.

1

The Apriori algorithm is based on the idea that for a pair o items to be frequent, each individual item should also be frequent. If the hamburguer-ketchup pair is frequent, the hamburger itself must also appear frequently in the baskets. The same can be said about the ketchup.

So for the algorithm, it is established a "threshold X" to define what is or it is not frequent. If an item appears more than X times, it is considered frequent.

The first step of the algorithm is to pass for each item in each basket, and calculate their frequency (count how many time it appears). This can be done with a hash of size N, where the position y of the hash, refers to the frequency of Y.

If item y has a frequency greater than X, it is said to be frequent.

In the second step of the algorithm, we iterate through the items again, computing the frequency of pairs in the baskets. The catch is that we compute only for items that are individually frequent. So if item y and item z are frequent on itselves, we then compute the frequency of the pair. This condition greatly reduces the pairs to compute, and the amount of memory taken.

Once this is calculated, the frequencies greater than the threshold are said frequent itemset.

Source: http://stackoverflow.com/questions/14164853/data-mining-and-frequent-datasets?rq=1

4 Data Mining Tips to Scrap Real Estate Data; Innovative Way to Give Realty Business a boost!

Internet has become a huge source of data – in fact; it has turned into a goldmine for the marketers, from where they can easily dig the useful data!

    Web scraping has become a norm in today’s competitive era, where one with maximum and relevant information wins the race!

Real Estate Data Extraction and Scraping Service

It has helped many industries to carve a niche in the market; especially real estate – Scraping real estate data has been of great help for professionals to reach out to a large number of people and gather reliable property data. However, there are some people for whom web scraping is still an alien concept; most probably because most of its advantages are not discussed.

    There are institutions, companies and organizations, entrepreneurs, as well as just normal citizens generating an extraordinary amount of information every day. Property information extraction can be effectively used to get an idea about the customer psyche and even generate valuable lead to further the business.

In addition to this, data mining has also some of following uses making it an indispensable part of marketing.

Gather Properties Details from Different Geographical Locations

You are an estate agent and want to expand your business to the neighboring city or state. But, then you are short of information. You are completely aware of the properties in the vicinity and in your town; however, with data mining services will help you to get an idea about the properties in the other state. You can also approach probable clients and increase your database to offer extensive services.

Online Offers and Discounts are just a Click Away

Now, it is tough to deal with the clients, show them the property of their choice and again act as a mediator between the buyer and seller. In all this, it becomes almost difficult to take a look at some special discounts or offers. With the data mining services, you can get an insight into these amazing offers. Thus, you can plan a move or even provide your client an amazing deal.

What people are talking about – Easy Monitoring of your Online Reputation

Internet has become a melting pot where different people come together. In fact, it provides a huge platform where people discuss about their likes and dislikes. When you dig into such online forums, you can get an idea of reputation that you or your firm holds. You can know what people think about you and where you require to buck up and where you need to slow down.

A Chance to Know your Competitors Better!

Last, but not the least, you can keep an eye on the competitor.  Real Estate is getting more competitive; and therefore, it is important to have knowledge about your competitors to get an upper hand. It will help you to plan your moves and strategize with more ease. Moreover, you also know what is that “something” that your competitor does not have and you have, with can be subtly highlighted.

Property information extraction can prove to be the most fruitful method to get a cutting edge in the industry.

Source: http://www.hitechbposervices.com/blog/4-data-mining-tips-to-scrap-real-estate-data-innovative-way-to-give-realty-business-a-boost/

Web Scraping for Fun & Profit

There’s a number of ways to retrieve data from a backend system within mobile projects. In an ideal world, everything would have a RESTful JSON API – but often, this isn’t the case.Sometimes, SOAP is the language of the backend. Sometimes, it’s some proprietary protocol which might not even be HTTP-based. Then, there’s scraping.

Retrieving information from web sites as a human is easy. The page communicates information using stylistic elements like headings, tables and lists – this is the communication protocol of the web. Machines retrieve information with a focus on structure rather than style, typically using communication protocols like XML or JSON. Web scraping attempts to bridge this human protocol into a machine-readable format like JSON. This is what we try to achieve with web scraping.

As a means of getting to data, it don’t get much worse than web scraping. Scrapers were often built with Regular Expressions to retrieve the data from the page. Difficult to craft, impossible to maintain, this means of retrieval was far from ideal. The risks are many – even the slightest layout change on a web page can upset scraper code, and break the entire integration. It’s a fragile means for building integrations, but sometimes it’s the only way.

Having built a scraper service recently, the most interesting observation for me is how far we’ve come from these “dark days”. Node.js, and the massive ecosystem of community built modules has done much to change how these scraper services are built.

Effectively Scraping Information

Websites are built on the Document Object Model, or DOM. This is a tree structure, which represents the information on a page.By interpreting the source of a website as a DOM, we can retrieve information much more reliably than using methods like regular expression matching. The most popular method of querying the DOM is using jQuery, which enables us to build powerful and maintainable queries for information. The JSDom Node module allows us to use a DOM-like structure in serverside code.

For purpose of Illustration, we’re going to scrape the blog page of FeedHenry’s website. I’ve built a small code snippet that retrieves the contents of the blog, and translates it into a JSON API. To find the queries I need to run, first I need to look at the HTML of the page. To do this, in Chrome, I right-click the element I’m looking to inspect on the page, and click “Inspect Element”.

Screen Shot 2014-09-30 at 10.44.38

Articles on the FeedHenry blog are a series of ‘div’ elements with the ‘.itemContainer’ class

Searching for a pattern in the HTML to query all blog post elements, we construct the `div.itemContainer` query. In jQuery, we can iterate over these using the .each method:

var posts = [];

$('div.itemContainer').each(function(index, item){

  // Make JSON objects of every post in here, pushing to the posts[] array

});

From there, we pick off the heading, author and post summary using a child selector on the original post, querying the relevant semantic elements:

    Post Title, using jQuery:

    $(item).find('h3').text()trim() // trim, because titles have white space either side

    Post Author, using jQuery:

    $(item).find('.catItemAuthor a').text()

    Post Body, using jQuery:

    $(item).find('p').text()

Adding some JSDom magic to our snippet, and pulling together the above two concept (iterating through posts, and picking off info from each post), we get this snippet:

var request = require('request'),

jsdom = require('jsdom');

jsdom.env(

  "http://www.feedhenry.com/category/blog",

  ["http://code.jquery.com/jquery.js"],

  function (errors, window) {

    var $ = window.$, // Alias jQUery

    posts = [];

    $('div.itemContainer').each(function(index, item){

      item = $(item); // make queryable in JQ

      posts.push({

        heading : item.find('h3').text().trim(),

        author : item.find('.catItemAuthor a').text(),

        teaser : item.find('p').text()

      });

    });

    console.log(posts);

  }

);

A note on building CSS Queries

As with styling web sites with CSS, building effective CSS queries is equally as important when building a scraper. It’s important to build queries that are not too specific, or likely to break when the structure of the page changes. Equally important is to pick a query that is not too general, and likely to select extra data from the page you don’t want to retrieve.

A neat trick for generating the relevant selector statement is to use Chrome’s “CSS Path” feature in the inspector. After finding the element in the inspector panel, right click, and select “Copy CSS Path”. This method is good for individual items, but for picking repeating patterns (like blog posts), this doesn’t work though. Often, the path it gives is much too specific, making for a fragile binding. Any changes to the page’s structure will break the query.

Making a Re-usable Scraping Service

Now that we’ve retrieved information from a web page, and made some JSON, let’s build a reusable API from this. We’re going to make a FeedHenry Blog Scraper service in FeedHenry3. For those of you not familiar with service creation, see this video walkthrough.

We’re going to start by creating a “new mBaaS Service”, rather than selecting one of the off-the-shelf services. To do this, we modify the application.js file of our service to include one route, /blog, which includes our code snippet from earlier:

// just boilerplate scraper setup

var mbaasApi = require('fh-mbaas-api'),

express = require('express'),

mbaasExpress = mbaasApi.mbaasExpress(),

cors = require('cors'),

request = require('request'),

jsdom = require('jsdom');

var app = express();

app.use(cors());

app.use('/sys', mbaasExpress.sys([]));

app.use('/mbaas', mbaasExpress.mbaas);

app.use(mbaasExpress.fhmiddleware());

// Our /blog scraper route

app.get('/blog', function(req, res, next){

  jsdom.env(

    "http://www.feedhenry.com/category/blog",

    ["http://code.jquery.com/jquery.js"],

    function (errors, window) {

      var $ = window.$, // Alias jQUery

      posts = [];

      $('div.itemContainer').each(function(index, item){

        item = $(item); // make queryable in JQ

        posts.push({

          heading : item.find('h3').text().trim(),

          author : item.find('.catItemAuthor a').text(),

          teaser : item.find('p').text()

        });

      });

      return res.json(posts);

    }

  );

});

app.use(mbaasExpress.errorHandler());

var port = process.env.FH_PORT || process.env.VCAP_APP_PORT || 8001;

var server = app.listen(port, function() {});

We’re also going to write some documentation for our service, so we (and other developers) can interact with it using the FeedHenry discovery console. We’re going to modify the README.md file to document what we’ve just done using API Blueprint documentation format:

# FeedHenry Blog Web Scraper

This is a feedhenry blog scraper service. It uses the `JSDom` and `request` modules to retrieve the contents of the FeedHenry developer blog, and parse the content using jQuery.

# Group Scraper API Group

# blog [/blog]

Blog Endpoint

## blog [GET]

Get blog posts endpoint, returns JSON data.

+ Response 200 (application/json)

    + Body

            [{ blog post}, { blog post}, { blog post}]

We can now try out the scraper service in the studio, and see the response:

Scraping – The Ultimate in API Creation?

Now that I’ve described some modern techniques for effectively scraping data from web sites, it’s time for some major caveats. First,  WordPress blogs like ours already have feeds and APIs available to developers - there’s no need to ever scrape any of this content. Web Scraping is not a replacement for an API. It should be used only as a last resort, after every endeavour to discover an API has already been made. Using a web scraper in a commercial setting requires much time set aside to maintain the queries, and an agreement with the source data is being scraped on to alert developers in the event the page changes structure.

With all this in mind, it can be a useful tool to iterate quickly on an integration when waiting for an API, or as a fun hack project.

Source: http://www.feedhenry.com/web-scraping-fun-profit/

Scraping websites using the Scraper extension for Chrome

If you are using Google Chrome there is a browser extension for scraping web pages. It’s called “Scraper” and it is easy to use. It will help you scrape a website’s content and upload the results to google docs.

Walkthrough: Scraping a website with the Scraper extension

•     Open Google Chrome and click on Chrome Web Store
•     Search for “Scraper” in extensions
•     The first search result is the “Scraper” extension
•     Click the add to chrome button.
•     Now let’s go back to the listing of UK MPs
•     Open http://www.parliament.uk/mps-lords-and-offices/mps/
•     Now mark the entry for one MP
•     http://farm9.staticflickr.com/8490/8264509932_6cc8802992_o_d.png
•     Right click and select “scrape similar…”
•     http://farm9.staticflickr.com/8200/8264509972_f3a9e5d8e8_o_d.png
•     A new window will appear – the scraper console
•     http://farm9.staticflickr.com/8073/8263440961_9b94e63d56_b_d.jpg
•     In the scraper console you will see the scraped content
•     Click on “Save to Google Docs…” to save the scraped content as a Google Spreadsheet.

Walkthrough: extended scraping with the Scraper extension

Note: Before beginning this recipe – you may find it useful to understand a bit about HTML. Read our HTML primer.

Easy wasn’t it? Now let’s do something a little more complicated. Let’s say we’re interested in the roles a specific actress played. The source for all kinds of data on this is the IMDB (You can also search on sites like DBpedia or Freebase for this kinds of information; however, we’ll stick to IMDB to show the principle)

    Let’s say we’re interested in creating a timeline with all the movies the Italian actress Asia Argento ever starred; where do we start?

    The IMDB has a quite comprehensive archive of actors. Asia Argento’s site is: http://www.imdb.com/name/nm0000782/

    If you open the page you’ll see all the roles she ever played, together with a title and the year – let’s scrape this information

    Try to scrape it like we did above

    You’ll see the list comes out garbled – this is because the list here is structured quite differently.

    Go to the scraper console. Notice the small box on the upper left, saying XPath?

    XPath is a query language for HTML and XML.

    XPath can help you find the elements in the page you’re interested in – all you need to do is find the right element and then write the xpath for it.

    Now let’s assemble our table.

    You’ll see that our current Xpath – the one including the whole information is “//div[3]/div[3]/div[2]/div”

    http://farm9.staticflickr.com/8344/8264510130_ae31697fde_o_d.png

    Xpath is very simple it tells the computer to look at the HTML document and select <div> element number 3, then in this the third one, the second one and then all <div> elements (which if you count down our list, results in exactly where you are right now.

•     However, we’d like to have the data separated out.
•     To do this use the columns part of the scraper console…
•     Let’s find our title first – look at the title using Inspect Element
•     http://farm9.staticflickr.com/8355/8263441157_b4672d01b2_o_d.png
•     See how the title is within a <b> tag? Let’s add the tag to our xpath.
•     The expression seems to work well: let’s make this our first column
•     In the “Columns” section, change the name of the first column to “title”
•     Now let’s add the XPATH for the title to it
•     The xpaths in the columns section are relative, that means “./b” will select the <b> element
•     add “./b” to the xpath for the title column and click “scrape”
•     http://farm9.staticflickr.com/8357/8263441315_42d6a8745d_o_d.png
•     See how you only get titles?
•     Now let’s continue for year? Years are within one <span>
•     Create a new column by clicking on the small plus next to your “title” column
•     Now create the “year” column with xpath “./span”
•     http://farm9.staticflickr.com/8347/8263441355_89f4315a78_o_d.png
•     Click on scrape and see how the year is added
•     See how easily we got information out of a less structured webpage?

Source: http://schoolofdata.org/handbook/recipes/scraper-extension-for-chrome/

Screenscraping from Java using jsoup – effective data gathering from websites

In a recent article I discussed screenscraping in a in hindsight fairly clumsy way (http://technology.amis.nl/blog/12786/building-java-object-graph-with-tour-de-france-results-using-screen-scraping-java-util-parser-and-assorted-facilities). While preparing for a series of articles on data visualizations, I had need of statistics regarding the Olympic Games – more specifically: the overall medal count per country during the 2008 Bejing Olympic Games. This information is readily available from dozens of websites. However, I could not find one hat offered the data in easy to process XML or CSV format – all websites had human consumers in mind.

Using screenscraping – we use a programmatic facility to consume the content that is intended to be displayed on screen to human users and subsequently process that content by extracting the required data from it. Some web-pages are easier to scrape than others – this depends on the richness of the HTML (the poorer the better for scraping), the required interactivity (JavaScript, AJAX – the less the better) and the structure used to present the data (tables, frequently despised by web developers, work rather well).

I came across a tool for screenscraping from Java, called jsoup – http://jsoup.org/. It turned out to be so incredibly easy to use – that I thouht I should share it.

Getting going with jsoup is as easy as can be:

1. download jsoup-1.6.1.jar (or whatever the latest version is) from http://jsoup.org/download

2. add this jar as a dependency in your project and/or application CLASSPATH

3. make use of jsoup in the code that does the screenscraping.

A simple example of code that uses jsoup (more examples on: http://jsoup.org/cookbook/):

One of the websites offering the overall medal count is http://www.databaseolympics.com/games/gamesyear.htm?g=26. The page looks as follows:

Image

Well, more importantly, the page looks like this:

Image

This means in terms of screenscraping: I will find the medal count for each country inside a TABLE element with styleclass pt8. Each country has a TR element. Only the first TR element does not represent a country score, as it is the table header. The first TD element in the TR represents the country. The name of the country can be retrieved as the text content from the A element in the TD. The next TD elements contain the numbers of medals in Gold, Silver, Bronze and Total.

The corresponding Java code with jsoup boils down to:

public static void main(String[] args) throws IOException, SQLException, InterruptedException {

        Document doc = Jsoup.connect(OlympicMedalMirrorProcessor.baseUrl + "?g=26").get();
        String title = doc.title();
        System.out.println(title);
        Element table = doc.select("table.pt8").get(0);
        Elements trs = table.select("tr");
        Iterator trIter = trs.iterator();
        boolean firstRow = true;
        while (trIter.hasNext()) {


            Element tr = (Element)trIter.next();
            if (firstRow) {
                firstRow = false;
                continue;
            }
            Elements tds = tr.select("td");
            Iterator tdIter = tds.iterator();
            int tdCount = 1;
            String country = null;
            Integer gold = null;
            Integer silver = null;
            Integer bronze = null;
            Integer total = null;
            // process new line
            while (tdIter.hasNext()) {

                Element td = (Element)tdIter.next();
                switch (tdCount++) {
                case 1:
                    country = td.select("a").text();
                    break;
                case 2:
                    gold = Integer.parseInt(td.text());
                    break;
                case 3:
                    silver = Integer.parseInt(td.text());
                    break;
                case 4:
                    bronze = Integer.parseInt(td.text());
                    break;
                case 5:
                    total = Integer.parseInt(td.text());
                    break;
                }

            }
            System.out.println(country + ": gold " + gold + " silver " + silver + " bronze " + bronze + " total " +
                               total);
        } //table rows

Source:http://technology.amis.nl/2011/08/03/screenscraping-from-java-using-jsoup-effective-data-gathering-from-websites/

Scraping Data: Site-specific Extractors vs. Generic Extractors

Scraping is becoming a rather mundane job with every other organization getting its feet wet with it for their own data gathering needs. There have been enough number of crawlers built – some open-sourced and others internal to organizations for in-house utilities. Although crawling might seem like a simple technique at the onset, doing this at a large-scale is the real deal. You need to have a distributed stack set up to take care of handling huge volumes of data, to provide data in a low-latency model and also to deal with fail-overs. This still is achievable after crossing the initial tech barrier and via continuous optimizations. (P.S. Not under-estimating this part because it still needs a team of Engineers monitoring the stats and scratching their heads at times).

Social Media Scraping

Focused crawls on a predefined list of sites

However, you bump into a completely new land if your goal is to generate clean and usable data sets from these crawls i.e. “extract” data in a format that your DB can process and aid in generating insights. There are 2 ways of tackling this:

a. site-specific extractors which give desired results

b. generic extractors that result in few surprises

Assuming you still do focused crawls on a predefined list of sites, let’s go over specific scenarios when you have to pick between the two-

1. Mass-scale crawls; high-level meta data - Use generic extractors when you have a large-scale crawling requirement on a continuous basis. Large-scale would mean having to crawl sites in the range of hundreds of thousands. Since the web is a jungle and no two sites share the same template, it would be impossible to write an extractor for each. However, you have to settle in with just the document-level information from such crawls like the URL, meta keywords, blog or news titles, author, date and article content which is still enough information to be happy with if your requirement is analyzing sentiment of the data.

cb1c0_one-size

A generic extractor case

Generic extractors don’t yield accurate results and often mess up the datasets deeming it unusable. Reason being

programatically distinguishing relevant data from irrelevant datasets is a challenge. For example, how would the extractor know to skip pages that have a list of blogs and only extract the ones with the complete article. Or delineating article content from the title on a blog page is not easy either.

To summarize, below is what to expect of a generic extractor.

Pros-

minimal manual intervention

low on effort and time

can work on any scale

Cons-

Data quality compromised

inaccurate and incomplete datasets

lesser details suited only for high-level analyses

Suited for gathering- blogs, forums, news

Uses- Sentiment Analysis, Brand Monitoring, Competitor Analysis, Social Media Monitoring.

2. Low/Mid scale crawls; detailed datasets - If precise extraction is the mandate, there’s no going away from site-specific extractors. But realistically this is do-able only if your scope of work is limited i.e. few hundred sites or less. Using site-specific extractors, you could extract as many number of fields from any nook or corner of the web pages. Most of the times, most pages on a website share similar templates. If not, they can still be accommodated for using site-specific extractors.

cutlery

Designing extractor for each website

Pros-

High data quality

Better data coverage on the site

Cons-

High on effort and time

Site structures keep changing from time to time and maintaining these requires a lot of monitoring and manual intervention

Only for limited scale

Suited for gathering - any data from any domain on any site be it product specifications and price details, reviews, blogs, forums, directories, ticket inventories, etc.

Uses- Data Analytics for E-commerce, Business Intelligence, Market Research, Sentiment Analysis

Conclusion

Quite obviously you need both such extractors handy to take care of various use cases. The only way generic extractors can work for detailed datasets is if everyone employs standard data formats on the web (Read our post on standard data formats here). However, given the internet penetration to the masses and the variety of things folks like to do on the web, this is being overly futuristic.

So while site-specific extractors are going to be around for quite some time, the challenge now is to tweak the generic ones to work better. At PromptCloud, we have added ML components to make them smarter and they have been working well for us so far.

What have your challenges been? Do drop in your comments.

Source: https://www.promptcloud.com/blog/scraping-data-site-specific-extractors-vs-generic-extractors/

'Scrapers' Dig Deep for Data on Web

At 1 a.m. on May 7, the website PatientsLikeMe.com noticed suspicious activity on its "Mood" discussion board. There, people exchange highly personal stories about their emotional disorders, ranging from bipolar disease to a desire to cut themselves.

It was a break-in. A new member of the site, using sophisticated software, was "scraping," or copying, every single message off PatientsLikeMe's private online forums.

Enlarge Image

Bilal Ahmed wrote about his health on a site that was scraped. Andrew Quilty for The Wall Street Journal.

PatientsLikeMe managed to block and identify the intruder: Nielsen Co., the privately held New York media-research firm. Nielsen monitors online "buzz" for clients, including major drug makers, which buy data gleaned from the Web to get insight from consumers about their products, Nielsen says.

"I felt totally violated," says Bilal Ahmed, a 33-year-old resident of Sydney, Australia, who used PatientsLikeMe to connect with other people suffering from depression. He used a pseudonym on the message boards, but his PatientsLikeMe profile linked to his blog, which contains his real name.

After PatientsLikeMe told users about the break-in, Mr. Ahmed deleted all his posts, plus a list of drugs he uses. "It was very disturbing to know that your information is being sold," he says. Nielsen says it no longer scrapes sites requiring an individual account for access, unless it has permission.

Related Reading

    Digits: Escaping the 'Scrapers'
    Complete Coverage: What They Know

Journal Community

The market for personal data about Internet users is booming, and in the vanguard is the practice of "scraping." Firms offer to harvest online conversations and collect personal details from social-networking sites, résumé sites and online forums where people might discuss their lives.

The emerging business of web scraping provides some of the raw material for a rapidly expanding data economy. Marketers spent $7.8 billion on online and offline data in 2009, according to the New York management consulting firm Winterberry Group LLC. Spending on data from online sources is set to more than double, to $840 million in 2012 from $410 million in 2009.

The Wall Street Journal's examination of scraping—a trade that involves personal information as well as many other types of data—is part of the newspaper's investigation into the business of tracking people's activities online and selling details about their behavior and personal interests.

Some companies collect personal information for detailed background reports on individuals, such as email addresses, cell numbers, photographs and posts on social-network sites.

Others offer what are known as listening services, which monitor in real time hundreds or thousands of news sources, blogs and websites to see what people are saying about specific products or topics.

One such service is offered by Dow Jones & Co., publisher of the Journal. Dow Jones collects data from the Web—which may include personal information contained in news articles and blog postings—that help corporate clients monitor how they are portrayed. It says it doesn't gather information from password-protected parts of sites.

It's rarely a coincidence when you see Web ads for products that match your interests. WSJ's Christina Tsuei explains how advertisers use cookies to track your online habits.

The competition for data is fierce. PatientsLikeMe also sells data about its users. PatientsLikeMe says the data it sells is anonymized, no names attached.

Nielsen spokesman Matt Anchin says the company's reports to its clients include publicly available information gleaned from the Internet, "so if someone decides to share personally identifiable information, it could be included."

Internet users often have little recourse if personally identifiable data is scraped: There is no national law requiring data companies to let people remove or change information about themselves, though some firms let users remove their profiles under certain circumstances.

California has a special protection for public officials, including politicians, sheriffs and district attorneys. It makes it easier for them to remove their home address and phone numbers from these databases, by filling out a special form stating they fear for their safety.

Data brokers long have scoured public records, such as real-estate transactions and courthouse documents, for information on individuals. Now, some are adding online information to people's profiles.

Many scrapers and data brokers argue that if information is available online, it is fair game, no matter how personal.

"Social networks are becoming the new public records," says Jim Adler, chief privacy officer of Intelius Inc., a leading paid people-search website. It offers services that include criminal background checks and "Date Check," which promises details about a prospective date for $14.95.

"This data is out there," Mr. Adler says. "If we don't bring it to the consumer's attention, someone else will."

Scraping for Your Real Name

PeekYou.com has applied for a patent for a way to, among other things, match people's real names to pseudonyms they use on blogs, Twitter and online forums.

Read PeekYou.com's patent application.

Enlarge Image

New York-based PeekYou LLC has applied for a patent for a method that, among other things, matches people's real names to the pseudonyms they use on blogs, Twitter and other social networks. PeekYou's people-search website offers records of about 250 million people, primarily in the U.S. and Canada.

PeekYou says it also is starting to work with listening services to help them learn more about the people whose conversations they are monitoring. It says it hands over only demographic information, not names or addresses.

Employers, too, are trying to figure out how to use such data to screen job candidates. It's tricky: Employers legally can't discriminate based on gender, race and other factors they may glean from social-media profiles.

One company that screens job applicants for employers, InfoCheckUSA LLC in Florida, began offering limited social-networking data—some of it scraped—to employers about a year ago. "It's slowly starting to grow," says Chris Dugger, national account manager. He says he's particularly interested in things like whether people are "talking about how they just ripped off their last employer."

Scrapers operate in a legal gray area. Internationally, anti-scraping laws vary. In the U.S., court rulings have been contradictory. "Scraping is ubiquitous, but questionable," says Eric Goldman, a law professor at Santa Clara University. "Everyone does it, but it's not totally clear that anyone is allowed to do it without permission."

Scrapers and listening companies say what they're doing is no different from what any person does when gathering information online—they just do it on a much larger scale.

"We take an incomprehensible amount of information and make it intelligent," says Chase McMichael, chief executive of InfiniGraph, a Palo Alto, Calif., "listening service" that helps companies understand the likes and dislikes of online customers.

Scraping services range from dirt cheap to custom-built. Some outfits, such as 80Legs.com in Texas, will scrape a million Web pages for $101. One Utah company, screen-scraper.com, offers do-it-yourself scraping software for free. The top listening services can charge hundreds of thousands of dollars to monitor and analyze Web discussions.

Some scrapers-for-hire don't ask clients many questions.

"If we don't think they're going to use it for illegal purposes—they often don't tell us what they're going to use it for—generally, we'll err on the side of doing it," says Todd Wilson, owner of screen-scraper.com, a 10-person firm in Provo, Utah, that operates out of a two-room office. It is one of at least three firms in a scenic area known locally as "Happy Valley" that specialize in scraping.

Enlarge Image

Some of the computer code behind screen-scraper.com's software. Chris Detrick for The Wall Street Journal

Screen-scraper charges between $1,500 and $10,000 for most jobs. The company says it's often hired to conduct "business intelligence," working for companies who want to scrape competitors' websites.

One recent assignment: A major insurance company wanted to scrape the names of agents working for competitors. Why? "We don't know," says Scott Wilson, the owner's brother and vice president of sales. Another job: attempting to scrape Facebook for a multi-level marketing company that wanted email addresses of users who "like" the firm's page—as well as their friends—so they all could be pitched products.

Scraping often is a cat-and-mouse game between websites, which try to protect their data, and the scrapers, who try to outfox their defenses. Scraping itself isn't difficult: Nearly any talented computer programmer can do it. But penetrating a site's defenses can be tough.

One defense familiar to most Internet users involves "captchas," the squiggly letters that many websites require people to type to prove they're human and not a scraping robot. Scrapers sometimes fight back with software that deciphers captchas.

More From the Series

    Web's New Goldmine: Your Secrets

    Personal Details Exposed Via Biggest Websites

    Microsoft Quashed Bid to Boost Web Privacy

    On Web's Cutting Edge, Anonymity in Name Only

    Stalking by Cellphone

    Google Agonizes Over Privacy

    The Tracking Ecosystem

    On the Web, Children Face Intensive Tracking

Some professional scrapers stage blitzkrieg raids, mounting around a dozen simultaneous attacks on a website to grab as much data as quickly as possible without being detected or crashing the site they're targeting.

Raids like these are on the rise. "Customers for whom we were regularly blocking about 1,000 to 2,000 scrapes a month are now seeing three times or in some cases 10 times as much scraping," says Marino Zini, managing director of Sentor Anti Scraping System. The company's Stockholm team blocks scrapers on behalf of website clients.

At Monster.com, the jobs website that stores résumés for tens of millions of individuals, fighting scrapers is a full-time job, "every minute of every day of every week," says Patrick Manzo, global chief privacy officer of Monster Worldwide Inc. Facebook, with its trove of personal data on some 500 million users, says it takes legal and technical steps to deter scraping.

At PatientsLikeMe, there are forums where people discuss experiences with AIDS, supranuclear palsy, depression, organ transplants, post-traumatic stress disorder and self-mutilation. These are supposed to be viewable only by members who have agreed not to scrape, and not by intruders such as Nielsen.

"It was a bad legacy practice that we don't do anymore," says Dave Hudson, who in June took over as chief executive of the Nielsen unit that scraped PatientsLikeMe in May. "It's something that we decided is not acceptable, and we stopped."

Mr. Hudson wouldn't say how often the practice occurred, and wouldn't identify its client.

The Nielsen unit that did the scraping is now part of a joint venture with McKinsey & Co. called NM Incite. It traces its roots to a Cincinnati company called Intelliseek that was founded in 1997. One of its most successful early businesses was scraping message boards to find mentions of brand names for corporate clients.

In 2001, the venture-capital arm of the Central Intelligence Agency, In-Q-Tel Inc., was among a group of investors that put $8 million into the business.

Intelliseek struggled to set boundaries in the new business of monitoring individual conversations online, says Sundar Kadayam, Intelliseek's co-founder. The firm decided it wouldn't be ethical to use automated software to log into private message boards to scrape them.

But, he says, Intelliseek occasionally would ask employees to do that kind of scraping if clients requested it. "The human being can just sign in as who they are," he says. "They don't have to be deceitful."

In 2006, Nielsen bought Intelliseek, which had revenue of more than $10 million and had just become profitable, Mr. Kadayam says. He left one year after the acquisition.

At the time, Nielsen, which provides television ratings and other media services, was looking to diversify into digital businesses. Nielsen combined Intelliseek with a New York startup it had bought called BuzzMetrics.

The new unit, Nielsen BuzzMetrics, quickly became a leader in the field of social-media monitoring. It collects data from 130 million blogs, 8,000 message boards, Twitter and social networks. It sells services such as "ThreatTracker," which alerts a company if its brand is being discussed in a negative light. Clients include more than a dozen of the biggest pharmaceutical companies, according to the company's marketing material.

Like many websites, PatientsLikeMe has software that detects unusual activity. On May 7, that software sounded an alarm about the "Mood" forum.

David Williams, the chief marketing officer, quickly determined that the "member" who had triggered the alert actually was an automated program scraping the forum. He shut down the account.

The next morning, the holder of that account e-mailed customer support to ask why the login and password weren't working. By the afternoon, PatientsLikeMe had located three other suspect accounts and shut them down. The site's investigators traced all of the accounts to Nielsen BuzzMetrics.

On May 18, PatientsLikeMe sent a cease-and-desist letter to Nielsen. Ten days later, Nielsen sent a letter agreeing to stop scraping. Nielsen says it was unable to remove the scraped data from its database, but a company spokesman later said Nielsen had found a way to quarantine the PatientsLikeMe data to prevent it from being included in its reports for clients.

PatientsLikeMe's president, Ben Heywood, disclosed the break-in to the site's 70,000 members in a blog post. He also reminded users that PatientsLikeMe also sells its data in an anonymous form, without attaching user's names to it. That sparked a lively debate on the site about the propriety of selling sensitive information. The company says most of the 350 responses to the blog post were supportive. But it says a total of 218 members quit.

In total, PatientsLikeMe estimates that the scraper obtained about 5% of the messages in the site's forums, primarily in "Mood" and "Multiple Sclerosis."

Source: http://online.wsj.com/articles/SB10001424052748703358504575544381288117888

My Experience in Choosing a Web Scraping Service

Recently I decided to outsource a web scraping project to another company. I typed “web scraping service” in Google, chose six services from the first two search result pages and sent the project specifications to all of them to get quotes. Eventually I decided to go another way and did not order the services, but my experience may be useful for others who want to entrust web scraping jobs to third party services.

If you are interested in price comparisons only and not ready to read the whole story just scroll down.

A list of web scraping services I sent my project to:

    www.datahen.com - Canadian web scraping service with nice web design
    webdata-scraping.com - Indian service by Keval Kothari
    www.iwebscraping.com - India based web scraping company (same as www.3idatascraping.com)
    scrapinghub.com - A scraping service founded by creators of Scrapy
    web-scraper.com - Yet another web scraping service
    grepsr.com - A scraping service that we already reviewed two years ago

Sending the request

All the services except scrapinghub.com have quite simple forms for the description of the project requirements. Basically, you just need to give your contact details and a project description in any form. Some of them are pretty (like datahen.com), some of them are more ascetic (like web-scraper.com), but all of them allow you to send your requirements to developers.

Scrapinghub.com has a quite long form, but most of the fields are optional and all the questions are quite natural. If you really know what you need, then it won’t be hard to answer all of them; moreover they rather help you to describe your need in detail.

Note, that in the context of the project I didn’t make a request for a scraper itself. I asked to receive data on a weekly basis only.

Getting responses

Since I sent my request on Sunday it would have been ok not to receive responses the same day, but I got the first response in 3 hrs! It was from web-scraper.com and stated that this project will cost me $250 monthly. Simple and clear. Thank you, Thang!

Right after that, I received the second response. This time it was Keval from webdata-scraping.com. He had some questions regarding the project. Then after two days he wrote me that it would be hard to scrape some of my data with the software he uses, and that he will try to use a custom scraper. After that he disappeared… ((

Then on Monday I received Cost & ETAT details from datahen.com. It looked quite professional and contained not only price, but also time estimation. They were ready to create such a scraper in 3-4 days for $249 and then maintain it for just $65/month.

On the same day I received a quote from iwebscraping.com. It was $60 per week. Everything is fine, but I’d like to mention that it wasn’t the last letter from them. After I replied to them (right after receiving the quote), I received a reminder letter from them every other day for about a week. So be ready for aggressive marketing if you ask them for a quote )).

Finally in two days after requesting a quote I got a response from scrapinghub.com. Paul Tremberth wrote me that they were ready to build a scraper for $1200 and then maintain it for $300/month.

It is interesting that I have never received an answer from grepsr.com! Two years ago it was the first web scraping service we faced on the web, but now they simply ignored my request! Or perhaps they didn’t receive it somehow? Anyway I had no time for investigation.

So what?

Let us put everything together. Out of six web scraping  services I received four quotes with the following prices:

Service     Setup fee     Monthly fee

web-scraper.com     -     $250
datahen.com     $249     $65
iwebscraping.com     -     $240
scrapinghub.com     $1200     $300

From this table you can see that  scrapinghub.com appears to be the most expensive service among those compared.

EDIT: These $300/month gives you as much support and development needed to fix a 5M multi-site web crawler, for example. If you need a cheaper solution you can use their Autoscraping tool, which is free, and would have costed around $2/month to crawl at my requested rates.

The average cost of monthly scraping is about $250, but from a long term perspective datahen.com may save you money due to their low monthly fee.

That’s it! If I had enough money available it would be interesting to compare all these services in operation and provide you a more complete report, but this is all I have for now.

If you have anything to share about your experience in using similar services, please contribute to this post by commenting on it below. Cheers!

Source: http://scraping.pro/choosing-web-scraping-service/

Why People Hesitate To Try Data Mining

What is hindering a number of people from venturing into the promising world of data mining? Despite so much encouragement, promotions, testimonials, and evidences of the benefits of online data collection, still only a handful take the challenge and really gain the pay offs it has to offer.

It may sound unthinkable that such an opportunity for success has been neglected by many. It may also sound absurd why many well-meaning individuals are hindered from enjoying the benefits of the blessings of the 21st century.

The Causes

After considerable observation and analysis of the human psyche, one can understand the underlying reasons behind the hesitance to try the profitable data mining service. The most common reasons why people are afraid to try new technology or why they remain passive and uninvolved are: fear; lack of knowledge; and pride.

Fear. The most paralyzing of human emotions is fear. It can, to some extent, cause a person to be insane, unprofitable, sick, and lost. Although fear is a normal reaction to certain stimuli and a natural feeling experienced by humans, it must always be monitored and controlled.  Usually, people share common fears, such as: fear of change; fear of anything new; and fear of the unknown.

Source:http://www.loginworks.com/blogs/web-scraping-blogs/people-hesitate-try-data-mining/

Why Web Scraping is Indispensable

The 21st century has opened the gates to hidden treasures and unlimited access to information globally without the constraints of time and space, through Internet technology. Along with this development comes the necessity for each business or company to get as much information as possible in order in order to thrive in the ever increasing demand for new innovations, comparisons, and trends.

Web scraping has consequently become an indispensable option to achieve all the needed data as quickly and efficiently as possible. In this view, data mining then appears to be the best and the only way to answer the present demand for updates, data, coping, foreknowledge, analysis, and evaluation. Indeed, information has inevitably become a valuable commodity and the most sought after product among online and offline entrepreneurs.

Need for Data

The increasing need for new data makes it possible for the experts to become increasingly creative in accessing information worldwide. The more knowledge one has, the better are his or her chances of growing and surviving. There seems to be no other time in the human existence where data has become so much a major source of revenue as the contemporary times.

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