This week’s blog put me in a bit of a difficult position because I more or less expended all my own thoughts about big data in last week’s blog post, and then Prof. Paul Moore goes and talks about more of the big data this week.
But anyway, we begun this week’s lecture by watching the TED talk The Birth of a Word by Deb roy. It’s about how he captured video of his son development from birth throughout the three years of his life using cameras and microphones installed in the ceilings of all the rooms in his home. Of course capturing his son’s development was the easy part, all of the footage then had to be analysed, he says in the talk that this amounted to 90,000 video and 140,000 hours of audio all of which added up to 200 terabytes of data. They used software to track movement of individuals as well as identify speech. From there they used this data to generate what he called ‘space time worms’, a way of visualising everyday life along the plane of time, something which found to be very interesting because I mentioned in last week’s blog about taking the data that people generate with the wearables and mobile devices and relating it to what they are actually doing in the real word. Roy has a very unique opportunity to visualise and learn from his very actions and habits when it comes to his interaction with his son.
Then they took it one step further and modified the fisheye images from the cameras and used projection mapping to overlay these images onto a 3D model of their home, effectively producing a ‘3D’ tourable video model of their environment although the video was still only 2D so the effect wasn’t fully realised I don’t think. On a side note this part in the talk really reminded me of the film Deja Vu with Denzel Washington where they discover a way to visualise the past in 3D and they use it to solve crimes.
After mapping out the videos from each room the analysed data could then be correlated with it in interesting ways, one of which was they took all of the words that the child learned and highlighted that word in all the transcripts. Because the software had not only transcribed all of the audio but also the location in the house that it had been said this data could be represented visually as what he called wordscapes. The one shown here is the the wordscape for ‘water’ and the place with the highest concentration of this word is in the kitchen as you might expect similarly the wordscape for ‘bye’ had the highest peaks around the front door and entrance hall.
Deb then went on to demonstrate how using these same kinds of data analysis techniques on other things like twitter and major news event uncovered valuable insights into audience interactions and social media influences. This included more 3D visualisations.
This graphic shows the social media users on top and the tv content below with all the lines representing the interactions between the two.
I think the research that Deb and his team are doing now at MIT is the precursor to how companies like IBM and google will look at the data everyone generates in the future.
Another valid point raised in class was to do with the internet of things and that was that you can’t really fully opt-out of it as it’s something that’s being built up as a kind of infrastructure around us all, the world is becoming more connected every day. I agree with all of this but am also interested to see if there will come a limit to how ‘integrated’ everyday people will want to be.
Google glass was also mentioned and this reminded me of a few videos by filmmaker and entrepreneur Casey Neistat because not only was the Explorer edition of google glass quite large and unsightly, earning early adopters the moniker of ‘glasshole’, it was quite impractical in that the display was an ever present prism in the top right of the wearer’s field of view. But apparently project has been reeled back in by the google lab dwellers with the promise that it hasn’t been abandoned just yet.
Plus a bonus google glass video by from Casey.