The flood of news from Google I/O continues as the company has announced a pair of services available to developers that provide public access to some of Google’s internal data-analysis tools. BigQuery, a service for analyzing massively large sets of data, and Prediction API, an interface for utilizing Google’s prediction algorithms, are now available to developers in the Google Code Labs. To break down these heavy new tools, we spoke with former Apple engineer and big-data geek Pete Warden.
Warden believes these new tools from Google could commoditize previously close-guarded technologies, allowing startups to quickly and easily leverage things like sentiment-analysis. “Assuming it does what it says on the label, this opens up a lot of technology problems to bootstrapped startups that previously required serious funding to tackle,” he told ReadWriteWeb.
One of the advantages to BigQuery, which can analyze billions of lines of data from trillions of records in mere seconds, is that it is written in a simple SQL-like language. This makes BigQuery a “user-friendly way to harness the power of massively distributed computing,” much like the Hive interface to Hadoop, said Warden.
“Both Hive and BigQuery are exciting because there’s millions of people who know SQL, though even a friendlier syntax doesn’t stop you from banging your head on the limitations of MapReduce,” he said.
But Warden seems more interested in the new Prediction API, which “promises to open up a lot of Google’s proprietary algorithms to startups,” he said.
“There’s a lot of meat in the list of examples they give, a lot of hard-to-solve problems that would previously have required an internal engineering team to tackle,” he said. “Imagine being able to get sentiment-analysis on all your customer emails at the touch of a button […] or integrating product recommendations into your small e-commerce site with almost no engineering effort.”
With the new Latitude API announced earlier today, startups have a whole new slew of developer toys from Google with which to play. Warden says with the new BigQuery and Prediction API, Google is providing access to these services on a silver platter.
“This is more than just Google opening up access to their infrastructure,” said Warden. “They’re offering something with intelligence about problem domains baked-in, which is much more valuable.”
