What framework in PHP can I use to write a web application (like peoplesmart.com, ancestry.com. spock.com) that will scrape or collect data from public records, social media about people?!

While I agree that PHP is not the ideal tool of choice for this task, for the record I totally disagree with ' reduction of PHP to being just "a web template language". If you're set on using PHP as you indicate, there's a built in library PHP-DOM (http://php.net/manual/en/book.dom.php) or this 3rd party library which makes the task syntactically easier called Simple HTML DOM (http://simplehtmldom.sourceforge.net/). As a suggestion, I know node.js has been used very successfully to build independent scraper robots which run lean and iterate quickly. I would consider making systems calls back to a node.js instance and then build your app in whatever language you prefer.

For the scraping part you definitely want to take a look at, Arc90's Readability. There's a PHP port too: http://www.keyvan.net/2010/08/php-readability/

For starters -- Give up on PHP for the application, PHP is not designed for doing this sort of task well. Alternatively, I suggest: Python -- look at Scrapy Java -- Nutch though hopefully the Droids framework will make it more useful Ruby -- Anemone Then you're just off to page parsing.