Friday, May 11, 2012

New ways of looking at the same thing.

Another one from the interesting file. It's from another statistics/data analysis site that I've been following recently. They reported on a project from the CS department at Carnegie Mellon that used social media data to look at the structure of neighborhoods. Various social media allow you use an app called foursquare that checks you in - basically it records your location.
 So they taken a whole bunch of data from a whole lot of people in a city, anonymized it so they not tracking recognizable individuals, it's the patterns people make not the individuals that's interesting in a study like this. And interesting patterns do emerge, as you can see. Each of these groups represent a community. The data has been tidied up, not everyone in each community only checks in at those points, a community is more an indication of a group of people who check in regularly at those sites. I see patterns like this and it makes me curious though. why are some of those communities spread out? Why are some so closely packed? Some communities are focused around parks, some appear to be focused around a very small set of streets. I haven't yet read the whole paper yet but its another a example of the spectacular, thought provoking patterns that a bit of statistics can pull out of everyday data.

No comments:

Post a Comment