Google Squared

Google has released two cool new products this week. First, there is a major update to the user interface for Street View.  The transitions between scenes are much smoother, and it reminds me of Microsoft’s Photosynth technology.  One gets the feeling of looking around corners, and actually being immersed in the environment.  For a good example, take a look at Times Square in New York.

As cool as this is, I’m even more excited about the public release of Google Squared, a new search product that creates tables for search results.   I had mourned the demise of Google Notebook, and haven’t really played around with Search Wiki, which is supposed to replace it. This new product is an excellent tool for research and comparison.

When you enter a search term a table is generated.  Google creates columns based on common features from all of the items, with the rows being entries for each item.  For example, if you entered the term “digital cameras” you would get a table like the one below:

The first three columns seem to be standard for every search – Item Name, Image, and Description.  Then Google adds columns for Manufacturer, Resolution, and Color Support.  In addition there is a blank column where you can add your own data.  Let’s say you were trying to do some comparison shopping for a camera.  You could type in “Price” in the column heading, and it would automatically fill in that data for each item, if available.

Here is another example using the search term “clouds“…

In this case columns have been created for Appearance, Precipitation, and Genus. These were automatically generated for the search term. However, only five items were returned for the list. I noticed that “Lenticular” was not on the list, so I added it manually. The information for the columns was populated, if available. As you can see in the image above, Google makes suggestions for other items that might fit the list.

As with any search engine, the level of specificity that you can provide will greatly improve the search results. I’m really showing my geeky side here, but I did a search for “starships” and got the following results…

The columns look right. There are listings for defenses, type, and armaments. Unfortunately the data in those columns makes no sense. The second item is for “Intrepid”, I’m assuming from the USS Intrepid on Star Trek. However, the image is of a Dodge Intrepid, hardly spaceworthy. The description is of the Intrepid Museum, a maritime museum for the naval vessel. For defenses, no value is given. Fortunately, there is a way to remove ambiguous results.

When I refined my search to “Federation Starships” I get better results…

All of the results are associated with Star Trek, and the columns have changed to reflect that context. Now we see Registry, Affiliation, and Class.

When you generate a search list, you can click on any of the boxes to get a link back to the original information. For example, in this butterfly search I can link to the original article on the University of Michigan’s website…

Just about any plural noun works well for a search as long as most of the more common search hits are non-ambiguous. A search for “pistols” shows the following…

…and a search for “Greenville SC BBQ” brings up a comparison listing of barbeque places in our area…

You can don’t even need to type in search terms.  You can start with an empty square and add your own columns and items.  Below I started a square by typing in several fictional detectives.  I was able to add columns for author and other information I needed…

 

If you have a Google account, you can save your searches, complete with any modification you have made, then return to them. Not only is it a very useful tool, but it’s highly addictive, as you start to create tables for all sorts of data.