How do you build an aggregator?

Ok, I've seen this announcement: -------- The service crawls 3,500 open source forges and gathers statistics and data on more than 300,000 open source projects and 300,000 open source developers. ---------<p>Or... --------- This services crawls and aggregates travel data from thousands of websites... ---------<p>And so on...<p>How actually an aggregator is built? How it works? Some insights would be helpful for me because I am trying to build a service which would be an aggregator (I won't build it myself but I want to understand).<p>Thank you.