Velocity Yahoo! has plugged its YQL web query language into a third-party API that lets developers access and analyze a sea of Twitter data dating back to 2006.
"This is the real strength of YQL," Yahoo! technical evangelist Tom Hughes-Croucher told The Reg this morning at the net-infrastructure-obsessed Velocity conference in Santa Clara, California. "It's not that YQL does everything. It's that we make it easier to use services that do the things you want."
YQL — the Yahoo! Query Language — is an umbrella API that lets app developers query, filter, and join data across disparate web services offered up by Yahoo! and the web at large. InfoChimps — yes, InfoChimps, a startup based on Austin, Texas — recently introduced a beta Query API that includes several calls for crunching old Twitter data, and Yahoo! has teamed with the startup to provide a Chimpified YQL interface.
InfoChimp offers, for instance, a call dubbed Trstrank, which uses an algorithm "similar to Google PageRank" that attempts to determine the influence of a particular Twitter user. "A developer could use YQL to generate search results from Twitter, pass them through a call...to order them...and reveal not just what's being said on Twitter, but what's being said by the Twitter big guns," InfoChimps' Sarah Nordquist said of the new YQL interface in a blog post.
As the name implies, YQL mimics MySQL. "Rather than thinking about specific databases, we're thinking about an SQL-like language that treats the internet as one big database," YQL product lead Jonathan Trevor previously told The Reg. But as Hughes-Croucher tells us, there are times when the analogy breaks down. "It's really like SQL, but there are a few differences because we're dealing with web services," he said. "We can't, say, do joins unless the web service supports it."
But the language can enable joins by plugging into services like InfoChimp's. According to Hughes-Croucher, InfoChimp has moved all sorts of data — including Twitter data and US census data — onto number-crunching Hadoop clusters, and then, with YQL, you can make calls that let you tap pre-crunched data. "They'll let you do queries without hosting the data yourself," he said, "and we give you an interface in order to plug into that system."
The YQL API is available as its own web service, and the service includes a web console here (sign-in required), where you can browse example queries and test your own. YQL now offers access to about 800 data tables, including data from Yahoo! services like Flickr, plus The New York Times, Facebook, Yelp, Microsoft Bing, and Twitter. ®