MIT’s Sociable Media Group is currently showcasing an interactive exhibit titled ‘Personas – how does the internet see you?’ – a search-engine-meets-data-visualisation that takes your first and last name, then scours the internet using ‘sophisticated natural language processing’ to produce a genome-type readout of how you are represented on the Internet.
Here’s mine (click for a full size version):
Trouble is that it only relies on first and last names, which means that in my case, it’s really just an amalgamation of how the internet sees anyone called Tom Gray.
This includes:
Tom Gray is a Mohawk Indian teenager who has moved to a new town with his mother and Tom Gray is widely considered one of the best bass players in bluegrass music.
If only.
So, pretty useless at showing how I’m represented on the web. But what it does showcase is some brilliant semantic searching: filtering words, phrases and concepts into broader categories like Sports, Fame, Fashion, Religious and the like.
The execution may be contrived, but the functionality behind it is an exciting one: search that doesn’t simply trawl through keywords, but rather ‘understands’ the content being searched, and categorises it accordingly.
Or, as one industry expert succintly put it: Find what I mean, not what I type.
Semantic search is the holy grail of search at the moment, and there are plenty of very clever people working on plenty of different ways of executing it at the moment. Pandia has more on five of the best semantic search engines.
Filed under: Uncategorized, data visualisation, MIT, search, semantic web, social media
Unscrambled 