How would you suggest starting data-mining for a project like this? Every time I think of a project like this, I convince myself that steam/amazon/twitter/Face

How would you suggest starting data-mining for a project like this? Every time I think of a project like this, I convince myself that steam/amazon/twitter/Facebook/etc will rate limit any bot that tries to scrape data from their service.