Tag Archives | Web Search

Here at Searchology, Google just announced something that will be available in Google Labs later this month. It’s called Google Squared, and I can’t quite tell if it’s going to be amazingly useful or just quirky and clever. It’s a search feature that returns results in a spreadsheet view, with information sorted into columns and rows. The demo involved doing a search on dog breeds (I’ll post images soon), with pictures of the breeds and information on factors like their size and energy level broken into fields.

Google is saying that the idea is a work in progress, and won’t always do what it’s supposed to–the demo also included a search for vegetables in which the search engine got confused and started populating the row for “squash” with information on the sport. But users can edit the results and save them–the notion is that a Google Squared result can be the starting point for big research jobs like choosing a dog.

It’ll be impossible to judge just how practical Google Squared is without doing a bunch of searches once it’s available, but the fact that Google is launching it at all is evidence of the semantic understanding of Web results that it’s gaining–the whole feature is dependent on the ability to turn unstructured text into highly regimented, database-like fields.

Also new today is a feature called Rich Snippets, which makes the little bits of text in results more useful–by identifying that a result is a review by a particular person, for instance, or extracting the location and profession of a person in a result so you can tell if it’s the person you’re looking for. And the last demo of the morning was for Google Sky Map, a virtual planetarium program for the Android platform that uses a phone’s accelerometer to let  you look in the sky and get a map of what you’re looking at.

Okay, some appallingly bad images of Google Squared in action (I shot these off a display at the Googleplex during a demo). Here’s a search for roller coasters, with a name column, one with an image, one with a description, one with the height and more:

Here’s a look at how Squared lets you see multiple items that might be appropriate for a cell in a square (which is what the spreadsheet-like views are called) and pick the best one. Note also that it shows where the data came from–and that Squared’s understanding of the data goes only so far (it has trouble distinguishing between a coaster’s height and the minimum height required to ride it):

Here’s a square of information about digital cameras (if Squared works well, it could be a potent research tool when you’re shopping for big-ticket items):

And here’s a square that a Google employee showed us when TechCrunch’s Mike Arrington asked to see results that didn’t work well (it’s for pizza, and shows restaurants in New York, Oakland, and Las Vegas):