Google Knowledge Graph – more info from Matt Cutts
Posted on August 16, 2012
So, Matt Cutts (some people call him an “SEO rockstar”, I call him “that guy that basically IS Google”) turned up for a suprise talk at SES San Fran. He talked about a few things that are worth noting:
Google Knowledge Graph
The Knowledge Graph is about connecting people, places and things and giving facts and data about the search you are doing. State of Search says:
Google figures out, based on your personalized data, what exactly you mean when you are looking for the Taj Mahal. But next to that they also show a box with results which could be about the other meanings of Taj Mahal.
Google studies the users and looks at relationships, not just in sites but all entities together. They are taking the social graph and the link graph and are combining the data from there together. In short: they are tracking our every move to figure out what kind of results we would like to have returned. Are you looking for (Indian) food or restaurants a lot? Your “Taj Mahal” search will show you more information on the Indian restaurants close to you, while if Google has figured out you love travel or culture you will see more about the monument.
From SES SF:
There are 3 billion connections with ‘real world things’. What he’s saying is that it is quite sophisticated already, but also very difficult. There are a lot of ‘Jason Smiths’ around. Knowledge Graph might be able to help and figure out what exactly fits your search best.
Later on in the talk Cutts talked a bit more about Knowledge Graph and how its set up. He says that Google Have built Knowledge Graph on Freebase – it’s Open Source, so you can look at it yourself (I’m assuming this is the URL - www.freebase.com)
Apart from Google Knowledge Graph, Matt Cutts also talked about:
Your emails in search results
I posted this the other day. Cutts says:
It might be scary for some people so we want people to opt-in.
Google & Tweets
Google doesn’t have access to the “Twitter firehose” any more. Cutts says:
It’s a lot more difficult for Google to see how many times specific pages are shared on Twitter.
…and went on to say:
If we could crawl Twitter in the full way we can, their infastructure wouldn’t be able to handle it.