PDF Scraping: Making Modern File Formats More Accessible

PDF Scraping: Making Modern File Formats More Accessible

Data scraping is the process of automatically sorting through information contained on the internet inside html, PDF or other documents and collecting relevant information to into databases and spreadsheets for later retrieval. On most websites, the text is easily and accessibly written in the source code but an increasing number of businesses are using Adobe PDF format (Portable Document Format: A format which can be viewed by the free Adobe Acrobat software on almost any operating system. See below for a link.). The advantage of PDF format is that the document looks exactly the same no matter which computer you view it from making it ideal for business forms, specification sheets, etc.; the disadvantage is that the text is converted into an image from which you often cannot easily copy and paste. PDF Scraping is the process of data scraping information contained in PDF files. To PDF scrape a PDF document, you must employ a more diverse set of tools.

There are two main types of PDF files: those built from a text file and those built from an image (likely scanned in). Adobe's own software is capable of PDF scraping from text-based PDF files but special tools are needed for PDF scraping text from image-based PDF files. The primary tool for PDF scraping is the OCR program. OCR, or Optical Character Recognition, programs scan a document for small pictures that they can separate into letters. These pictures are then compared to actual letters and if matches are found, the letters are copied into a file. OCR programs can perform PDF scraping of image-based PDF files quite accurately but they are not perfect.

Once the OCR program or Adobe program has finished PDF scraping a document, you can search through the data to find the parts you are most interested in. This information can then be stored into your favorite database or spreadsheet program. Some PDF scraping programs can sort the data into databases and/or spreadsheets automatically making your job that much easier.

Quite often you will not find a PDF scraping program that will obtain exactly the data you want without customization. Surprisingly a search on Google only turned up one business, (the amusingly named ScrapeGoat.com that will create a customized PDF scraping utility for your project. A handful of off the shelf utilities claim to be customizable, but seem to require a bit of programming knowledge and time commitment to use effectively. Obtaining the data yourself with one of these tools may be possible but will likely prove quite tedious and time consuming. It may be advisable to contract a company that specializes in PDF scraping to do it for you quickly and professionally.

Let's explore some real world examples of the uses of PDF scraping technology. A group at Cornell University wanted to improve a database of technical documents in PDF format by taking the old PDF file where the links and references were just images of text and changing the links and references into working clickable links thus making the database easy to navigate and cross-reference. They employed a PDF scraping utility to deconstruct the PDF files and figure out where the links were. They then could create a simple script to re-create the PDF files with working links replacing the old text image.

A computer hardware vendor wanted to display specifications data for his hardware on his website. He hired a company to perform PDF scraping of the hardware documentation on the manufacturers' website and save the PDF scraped data into a database he could use to update his webpage automatically.

PDF Scraping is just collecting information that is available on the public internet. PDF Scraping does not violate copyright laws.

PDF Scraping is a great new technology that can significantly reduce your workload if it involves retrieving information from PDF files. Applications exist that can help you with smaller, easier PDF Scraping projects but companies exist that will create custom applications for larger or more intricate PDF Scraping jobs.

Source: http://ezinearticles.com/?PDF-Scraping:-Making-Modern-File-Formats-More-Accessible&id=193321

Scrape amazon and price your product the right way – A use case

Scrape amazon and price your product the right way – A use case

So you built a product that you want to sell through Amazon.

How do you price your product? 

Amazon is the world’s largest online retailer. Millions of products are sold through amazon.  a lot of people make their living selling through Amazon. One of the biggest mistake people do in Amazon is that they price their product the wrong way. Sometimes they sell overpriced products, sometimes they sell the underpriced product. Both situations are toxic for the business.

We recently worked with a company that helps small businesses sell the products efficiently through amazon and other marketplaces. One of the key things they are doing is helping people with pricing their product the right way.

What I learned from them is that price is a relative term and a lot of people does not understand it. Pricing is a function of the positioning of  your product in the market.

We need to collect the data using  a technique called web scraping to understand how to position the product. You can get the  data in a CSV file which can be used for analysis.

1) What is the average price of a comparable product?

Understanding the pricing  strategy of your competitors products  is the first step in solving the problem. This can give you a range in which you can price your product. You can get the pricing data by scraping amazon

2) Is this a premium product?

People always pay a premium price for a premium product. What makes a product premium? – A product is considered premium only when the customer believe it is worth the price. Excellent marketing and branding are the ways to position your product as a premium product. You can get the relevant data by scraping amazon.

3) What are the problems with your competitor products?

Your competitor products might be having some defects. Or they might not be addressing a relevant problem. You have every chance of success If you are solving a problem that your competitor doesn’t. You can find these problems by analyzing the product reviews of your competitors. You can get review data by scraping amazon.

By analyzing data you can reach at a point where your profit margin looks healthy and pricing looks sensible. Buyers buy the value, not your product. Differentiate your product and position it as a superior product. Give people a reason to buy and that is the only way to succeed.

Source: http://blog.datahut.co/scrape-amazon-and-price-your-product-the-right-way-a-use-case/

Why Outsourcing Data Mining Services?

Why Outsourcing Data Mining Services?

Are huge volumes of raw data waiting to be converted into information that you can use? Your organization's hunt for valuable information ends with valuable data mining, which can help to bring more accuracy and clarity in decision making process.

Nowadays world is information hungry and with Internet offering flexible communication, there is remarkable flow of data. It is significant to make the data available in a readily workable format where it can be of great help to your business. Then filtered data is of considerable use to the organization and efficient this services to increase profits, smooth work flow and ameliorating overall risks.

Data mining is a process that engages sorting through vast amounts of data and seeking out the pertinent information. Most of the instance data mining is conducted by professional, business organizations and financial analysts, although there are many growing fields that are finding the benefits of using in their business.

Data mining is helpful in every decision to make it quick and feasible. The information obtained by it is used for several applications for decision-making relating to direct marketing, e-commerce, customer relationship management, healthcare, scientific tests, telecommunications, financial services and utilities.

Data mining services include:

•     Congregation data from websites into excel database
•     Searching & collecting contact information from websites
•     Using software to extract data from websites
•     Extracting and summarizing stories from news sources
•     Gathering information about competitors business

In this globalization era, handling your important data is becoming a headache for many business verticals. Then outsourcing is profitable option for your business. Since all projects are customized to suit the exact needs of the customer, huge savings in terms of time, money and infrastructure can be realized.

Advantages of Outsourcing Data Mining Services:

•     Skilled and qualified technical staff who are proficient in English
•     Improved technology scalability
•     Advanced infrastructure resources
•     Quick turnaround time
•     Cost-effective prices
•     Secure Network systems to ensure data safety
•     Increased market coverage

Outsourcing will help you to focus on your core business operations and thus improve overall productivity. So data mining outsourcing is become wise choice for business. Outsourcing of this services helps businesses to manage their data effectively, which in turn enable them to achieve higher profits.

Source: http://ezinearticles.com/?Why-Outsourcing-Data-Mining-Services?&id=3066061

Scraping Yelp Data and How to use?

Scraping Yelp Data and How to use?

We get a lot of requests to scrape data from Yelp. These requests come in on a daily basis, sometimes several times a day. At the same time we have not seen a good business case for a commercial project with scraping Yelp.

We have decided to release a simple example Yelp robot which anyone can run on Chrome inside your computer, tune to your own requirements and collect some data. With this robot you can save business contact information like address, postal code, telephone numbers, website addresses etc.  Robot is placed in our Demo space on Web Robots portal for anyone to use, just sign up, find the robot and use it.

How to use it:

    Sign in to our portal here.
    Download our scraping extension from here.
    Find robot named Yelp_us_demo in the dropdown.
    Modify start URL to the first page of your search results. For example: http://www.yelp.com/search?find_desc=Restaurants&find_loc=Arlington,+VA,+USA
    Click Run.
    Let robot finish it’s job and download data from portal.

Some things to consider:

This robot is placed in our Demo space – therefore it is accessible to anyone. Anyone will be able to modify and run it, anyone will be able to download collected data. Robot’s code may be edited by someone else, but you can always restore it from sample code below. Yelp limits number of search results, so do not expect to scrape more results than you would normally see by search.

In case you want to create your own version of such robot, here it’s full code:

// starting URL above must be the first page of search results.
// Example: http://www.yelp.com/search?find_desc=Restaurants&find_loc=Arlington,+VA,+USA

steps.start = function () {

   var rows = [];

   $(".biz-listing-large").each (function (i,v) {
     if ($("h3 a", v).length > 0)
       {
        var row = {};
        row.company = $(".biz-name", v).text().trim();
        row.reviews =$(".review-count", v).text().trim();
        row.companyLink = $(".biz-name", v)[0].href;
        row.location = $(".secondary-attributes address", v).text().trim();
        row.phone = $(".biz-phone", v).text().trim();
        rows.push (row);
      }
   });

   emit ("yelp", rows);
   if ($(".next").length === 1) {
     next ($(".next")[0].href, "start");
   }
 done();
};

Source: https://webrobots.io/scraping-yelp-data/

How to use Web Content Extractor(WCE) as Email Scraper?

How to use Web Content Extractor(WCE) as Email Scraper?

Web Content Extractor is a great web scraping software developed by Newprosoft Team. The software has easy to use project wizard to create a scraping configuration and scrape data from websites.

One day I came to see the Visual Email Extractor which is also product of Newprosoft and similar to Web Content Extractor but it’s primary use is to scrape email addresses by crawling websites you feed to the scraper. I had noticed that with the little modification in Web Content Extractor project configuration you can use it same as Visual Email Extractor to extract email addresses.

In this post I will show you what configuration makes the Web Content Extractor to extract email addresses. I still recommend Visual Email Extractor as it has lot more features then extracting email using WCE.

Here are the configuration that makes WCE to Extract Emails.

Step 1 : Open Web Content Extractor and Create New Project and Click on Next.

Step 2:  Under Crawling Rules -> Advanced Rules Tab do the following settings

Crawling Level 1 Settings

Follow Links if link text equals:
*contact*; *feedback*; *support*; *about*

for 'Follow Links if link text equals' text box enter following values:
contact; feedback; support; about

for 'Do not Follow links if URL contains' text box enter following values:

google.; yahoo.; bing; msn.; altavista.; myspace.com; youtube.com; googleusercontent.com; =http; .jpg; .gif; .png; .bmp; .exe; .zip; .pdf;

Set 'Maximum Crawling Deapth' to 2

set 'Crawling Order' to Deapth First Crawling

Tick mark below below check boxes:

->Follow all internal links

  Crawling Level 2  Settings

set 'Follow links if link text equals' to below value

*contact*; *feedback*; *support*; *about*

set 'Follow links if url contains' text box to below value

contact; feedback; support; about

set 'DO NOT follow links if url contains' text box to below value

=http

Step 3 After doing above settings now click on Next  -> in Extraction Pattern window -> Click on Define ->  in Web Page Address (URL) give any URL where email is given.  and click on  + sign right of Date Fields to define scraping pattern.

Now inside HTML Structure selects HTML check box or Body check box which means for each page it will take whole page content to parse data.

Now last settings to extract emails from page using regular expression based email extraction function.  Open Predefined Script window and select ‘Extract_Email_Addresses‘ and click on OK. and if you have used page that contains email then in Script Result’ you will be able to see the harvested email.

Hope this will help you to use your Web Content Extractor as a Email Scraper.. Share your view in comment.

Source: http://webdata-scraping.com/use-web-content-extractor-as-email-scraper/

Run Code Template – New Feature Added to Fminer Web Scraping Tool

Run Code Template – New Feature Added to Fminer Web Scraping Tool

Fminer is one of the powerful web scraping software, I already given brief of all the Fminer features in previous post. In this post I am going to introduce one of the interesting feature of fminer which is Run Code Template that is recently added to Fminer, this feature is similar to “Fminer Run Code” action but it’s different in a way you can use it. The Run Code Action you can use inside the data scraping flow and python code get executed when scraper start running.

While Run Code Templates are the saved python code snippets that you can run on the data tables after scraping completes. Assume if you get white space in scraped data then you can easily trim this left and right spaces by just executing “strip_column” template, see the code of that template below.

'''Strip all data of a column in data table
Remove the blank of data in the head and the tail.
'''

tabName = '[%table1|data table%]'
colName = '[%table1.column1|table column for strip%]'

tab = tables[tabName]
for i, row in enumerate(tab):
    row[colName] = row[colName].strip()   
    tab.edit_row(i, row)

This template comes with Fminer and few other template like “merge_tables_with_same_columns”.  Below are the steps how you can execute template python code on scraped data.

Step 1: Click on second icon from right that says “Run Code” under the Data section

Step 2: One popup will appear, you need to click on “Templates” icon and choose the template you want to execute and then click on Ok.

Step 3: Now the window will appear for configuration that will ask you to choose the table and column under that table on which you want to execute the code. Now click on Ok again.

Step 4: Now you can see the code of that template, now you can click on execute icon and script will start running, based on number of records it will take time to finish execution.

In many web scraping projects I found this template code very handy for cleaning data and making life easy. Templates are stored at following path so you can create your own template with customized code.

C:\Program Files (x86)\FMiner\templates

I have created one template which I use to remove HTML code that comes while scraping badly organized HTML pages. Below is the code of template for stripping html:

'''Strip HTML will remove all html tags of a column in data table.
'''
import re
tabName = '[%table1|data table%]'
colName = '[%table1.column1|table column for substring%]'
colNew = '[%table1.column1|table column to add new data%]'
tab = tables[tabName]
for i, row in enumerate(tab):
    cleanr =re.compile('<.*?>')
    cleantext = re.sub(cleanr,'', row[colName])
    row[colNew] = cleantext 
    tab.edit_row(i, row)

Stay connected as I am going to post more code templates that will make your web scraping life easy and manipulate data on fly.

Source: http://webdata-scraping.com/run-code-template-new-feature-added-fminer-web-scraping-tool/

Benefits of Ruby over Python & R for Web Scraping

Benefits of Ruby over Python & R for Web Scraping

In this data driven world, you need to be constantly vigilant, as information and key data for an organization keeps changing all the while. If you get the right data at the right time in an efficient manner, you can stay ahead of competition. Hence, web scraping is an essential way of getting the right data. This data is crucial for many organizations, and scraping technique will help them keep an eye on the data and get the information that will benefit them further.

Web scraping involves both crawling the web for data and extracting the data from the page. There are several languages which programmers prefer for web scraping, the top ones are Ruby, Python & R. Each language has its own pros and cons over the other, but if you want the best results and a smooth flow, Ruby is what you should be looking for.

Ruby is very good at production deployments and using Ruby, Redis & Chef have proven to be a great combination. String manipulation in Ruby is very easy because it is based on Perl syntax. Also, Ruby is great for analyzing web pages using  one of the very powerful gems called Nokogiri. Nokogiri is much easier to use as compared to other packages and libraries used by R and Python respectively. Nokogiri can deal with broken HTML / HTML fragments easily. Ruby also has many extensions, such as Sanitize and Loofah, that can help clean up broken HTML.

Python programmers widely use a library called Beautiful Soup for pulling data out of HTML & XML files. It works with your favorite parser to provide idiomatic ways of navigating, searching, and modifying the parse tree. It commonly saves programmers hours or days of work. R programmers have a new package called rvest that makes it easy to scrape data from html web pages, by libraries like beautiful soup. It is designed to work with magrittr so that you can express complex operations as elegant pipelines composed of simple, easily understood pieces.

To help you understand it more effectively, below is a comprehensive infographic for the same.

Ruby is far ahead of Python & R for cloud development and deployments.  The Ruby Bundler system is just great for managing and deploying packages from Github. Using Chef, you can start up and tear down nodes on EC2, at will, and monitor for failures,  scale up or down, reset your IP addresses, etc. Ruby also has great testing frameworks like Fakeweb and Capybara, making it almost trivial to build a great suite of unit tests and to include advanced features, like crawling  and scraping using webkit / selenium. 

The only disadvantage to Ruby is lack of machine learning and NLP toolkits, making it much harder to emulate the capacity of a tool like Pattern.  It can still be done, however, since most of the heavy lifting can be done asynchronously using Unix tools like liblinear or vowpal wabbit.

Conclusion

Each language has its plus point and you can pick the one which you are most comfortable with. But if you are looking for smooth web scraping experience, then Ruby is the best option. That has been our choice too for years at PromptCloud for the best web scraping results. If you have any further questions about this, then feel free to get in touch with us.

Source: https://www.promptcloud.com/blog/benefits-of-ruby-for-web-scraping