Almost Scraping: Web Scraping for Non-Programmers Michelle Minkoff, PBSNews.org Matt Wynn, Omaha World-Herald
What is Web scraping?
The *all-knowing* Wikipedia says:
“ Web scraping (also called Web harvesting or Web data extraction ) is a computer software technique of extracting information from websites. …Web scraping focuses more on the transformation of unstructured Web content, typically in HTML format, into structured data that can be stored and analyzed in a central local database or spreadsheet. Web scraping is also related to Web automation, which simulates human Web browsing using computer software. Uses of Web scraping include online price comparison, weather data monitoring, website change detection, Web research, Web content mashup and Web data integration.”
Why do I want to Web scrape?
Journalists like to find stories
Editors like stories that are exclusive
Downloading a dataset is like going to a press conference, anyone can grab and use it.
Web scraping is like an enterprise story, less likely to be picked up by all.
Puts more control back into your hands
What kind of data can I get?
Laws (Summary of same-sex marriage laws for each state, pdfs)
Photos (pictures associated with all players on a team you’re highlighting, all mayoral candidates)
Recipe ingredients (NYT story about peanut butter)
Health care (see ProPublica’s Dollars for Docs project)
Links, images, dates, names, categories, tags, anything with some sort of repeatable structure
Yahoo Pipes http://pipes.yahoo.com/pipes
Access and manipulate RSS feeds, which are often a flurry of information
Sort, filter, combine your information
Format that info to fit your needs (date formatter)
Pair with Versionista, which can create an RSS feed of changes to a Web site to keep tabs on what’s changing. This was done to great effect by ProPublica’s team in late 2009, esp. by Scott Klein and then-intern Brian Boyer, now at Chicago Tribune
For sites that follow a repetitive formula spanning multiple pages, like index pg & detail page, maybe with a search results page in the middle
Like a good employee, train it once, and then let it churn.
Query, select and filter your data in the Web app, then export in format of your choice.
Can check your data and stay up-to-date on your data set
Will go more in depth on Needle in Sat.’s hands-on lab at 10 a.m.
Record repetitive tasks that you do every day, and keep them as a data set
Think of it like a bookmark, but if you could include logging in, or entering a search term, as part of that bookmark
Useful for stats you check every day, scores for your local sports team, stocks if you’re a biz reporter, etc.
More complex function allows you to extract multiple data points on a page, like from an HTML table.
Versatile Firefox extension
Can use it for certain defaults (links, images)
Dig through the HTML hierarchy tree
Structural elements (<h3>)
Stylistic elements (<strong>)
Download list of attached files or files themselves
More options if you buy Pro version
Will discuss in-depth and use in hands-on lab on Saturday at 10 am
Non-programming scrapers can’t do everything, but have the power to get you started. Some say “Program or be programmed,” but this is a compromise.
Legal permissions still apply, so don’t use scraped info you don’t have the right to.
Something to consider. How does this apply to what you do every day, and how scraping could contribute to your job?
“ The businesses that win will be those that understand how to build value from data from wherever it comes. Information isn’t power. The right information is.” – media consultant Neil Perkin wrote in Marketing Week