Last week Google announced that it had created a new kind of search called a “knowledge graph.” Lifehacker‘s article said it would “bring[s] smarter semantic results” because it “connects your search query to Google’s knowledgebase of over 500 million people, things, and places to show you relevant info in a sidebar along your search results.”
So, what does that mean to you? It means Google just got more useful (and it was already pretty useful!) Google searches should bring you more information of different types. Let me give you an example and then explain the jargon that is being thrown around to explain this idea.
Suppose you were interested in “roses.” You would most likely go to Google and search on “roses.” In the past, Google would bring back webpages where the letters “roses” appeared (in that order) or which had descriptions that it had something to do with “roses.” That might mean poems with roses, the War of the Roses, the musical group Guns and Roses, the flower roses, or people who have first names of Roses (probably in error), or other places where the letters “roses” appeared. The term “roses” did not mean anything to Google; the search engine just found places where the letters were used.
Now when you type in “roses,” the new and improved Google “understands” that those letters in that order refer to something, a flower. So, in addition to the searches that you might have gotten elsewhere, you will now get a side panel that defines what roses are and how they are classified. You might also get a list of places at which to purchase roses (because that’s what one does to obtain them). It also takes advantage of what other databases Google has that mention roses and what other users have found useful when searching for roses.
Why didn’t they just say that Google got better? This is one of those examples where computer people did just say that Google got better, but they did it with a lot of jargon. “Semantic” refers to understood meaning; in this case that the search engine behaves as if it understands the meaning of the string of letters “roses” and its relationship with other things (like stores and gardens and bouquets). It is no longer just finding that word, but is now looking for information about the flower, roses. It is as if you had a librarian there helping you with your search.
So, what is all this about the graph? There is no graph on the search page. Again, it is jargon. To computer people, the “graph” refers to relationships among things. So it refers to the way that Google is now connecting its databases and its relationships among pages. It is making those connections to make the searchers more meaningful.
In Google’s announcement, they provided the search example “da Vinci.” This provides a nice example of the new search results. The screen you see is shown below.
As you can see from the screen shot above, the left side of the screen provides the typical Google search results. It provides not only information about the artist, but also information about syrups, wedding dresses and surgical procedures that share the name. The part that is different is the information on the right. Since most people who search for da Vinci are looking for the artist, they provide information about him, taken from a variety of databases, for easy access.
How well this better informed search will work for you depends on a few things. First, of course, it depends on how similar your searches are to those of other users. The closer you are to “typical” searchers, the more likely you will get this enhanced information. Second, it depends on how well you scope your search. If you are very good and tend to give Google a precise set of search terms, you most likely get to the information very quickly and this will not help much. If, instead, you are terribly general in your search terms, you may or may not get this information. However, if your terms generally provide some boundary, and you are searching for items for which Google has a good network, you will probably find this enhanced search is useful for you.
As an example, earlier this evening, I conducted a search on “Paul Gray,” who is a long standing colleague and friend who recently passed away. This search generated the enhanced results, but apparently there is a musician named Paul Gray and there was much information about him. Most people who search on Google probably want information about him. However, Google was smart enough to know there was another Paul Gray that might be of interest to a subset of searchers. So, at the bottom there was another box with information:
Google can understand that “Paul Gray” is a person’s name and there are two of them. That’s pretty cool!