News for hackers. Stuff that matters.
Re:The hell that is a restaurant website by rubyrescue on Tuesday February 16, 2010 @10:40PM (#1130780)
we scraped a huge amount of data, then hired people to clean it up. it does scale better down in south america where wages are quite a bit cheaper but it doesn't scale enough. i would say the difference is i can hire someone here to do some data gathering and quality control on-site whereas i'd have to outsource to mturk and get lower quality if we were based in the states.

however, since we're converging on a model that's slightly further toward the 'curated' side of the spectrum than a loopt/gowalla/nextstop, i think we're OK with doing some manual editing of things like open hours, and leaving the rest to the community - all of our community editing pieces aren't out yet (we all still have full-time jobs and/or ambitious clients)...

I love the check-in statistical analysis concept for data generation or at least quality control. very cool for higher volume locations.

Re:The hell that is a restaurant website by pyman on Wednesday February 17, 2010 @04:27AM (#1131207)
How do you scrape if they are all flash sites!
Re:The hell that is a restaurant website by jmtame on Thursday February 18, 2010 @09:49PM (#1136022)
you don't scrape the sites individually. there are large directories that exist already.