Sunday, September 23, 2012

Wolfram Alpha's Social Network Analysis

A couple of weeks ago I ran across this new feature of Wolfram Alpha designed to analyze your Facebook activity. I think this is an excellent example of how to summarize and elegantly display some of the mass of social information being captured over the net. We are starting to get enough data in the social sciences that hopefully the soft sciences can be made a bit harder. We're a long ways off from what I want to do with that, but are making some progress. That is one of the reasons I liked Klout, despite the rather unfortunate taste it leaves in some people's mouths. It takes a massively complex set of data and tries to reduce it down to something easily measurable so that people can act on it.

The new Wolfram Alpha project does a good job of that. It is on a very small scale, but I am able to look at my own analysis and learn some useful information. I can see at a glance the various friend groups I'm a part of, where my friends come from, who I'm interacting with over Facebook, what time of day this interaction is occurring, broad statistics of the religions and relationship statuses of my friends, and so forth. I find it very informative.

We still have got a long long ways to go. Get Wolfram Alpha's Facebook analysis expanded to all the social networks Klout covers, and cross-reference all of that data. Have location aware services such as Google Latitude and Foursquare tied in and watch the daily flow of people. Add more economic measurements and get social network systems monitoring how money flows through society. It would be great if someone could find a way to convince people to make public, or least public after being anonymized, data from sources such as Mint.com. Of course, I didn't use Mint.com or any such system until I started writing this article. Getting enough people to plug financial measurements into social networking software for a better understanding of economics is a ways off, but perhaps that's an open market ripe for invention.

I'd give a lot to have better economic theories backed by data to inform our politicians. True, politicians don't often feel the need to back their actions with data, but there's such a massive amount of public ignorance that any tool to help would be of benefit. One study that comes to mind was asking Americans what their desired level of wealth inequality in society was. The results showed that Americans greatly underestimated wealth inequality in the US, and their desired wealth distribution was similar to what Sweden has. People, myself included, tend to have a very poor idea of the construction of society, despite the massive amount of data out there. We still need better systems for working with and displaying this data.


I want, perhaps not a Newtonian revolution that reduces the messiness of the social sciences to clean lines and equations, humans are too complex for that to be possible, but I at least want something better than we have now. If not ways to predict the future, at least better ways of visualizing what is going on around us.  Now, I obviously don't think that having better data visualization will erase human ignorance, but it has to help. Making it easier for everyday people to better see what is going on can't hurt. It feels like there's a bunch of analysis being done inside companies to improve marketing or make financial predictions, and not much progress in creating broad, useful scientific theories. Perhaps I'm just not up to date on what is being learned, ignorantly dreaming of revolutionary theories, but I take comfort in things like Wolfram Alpha doing a bit to peel back the curtain.

1 comment:

  1. Wow Matt, not only is this *extremely* cool, but parts of it remind me of a very similar network on El Dorado's wall...

    ReplyDelete