From Mother Earth News to SWAT in 9 Steps

Amazon’s “Customers who bought X also bought Y” function displays thousands of paths through hundreds of demographic niches of America’s magazines. You don’t need Amazon to tell you that readers of ‘Guns & Ammo’ are more likely to buy ‘Handguns’ than ‘PETA News.’ But what Amazon can tell you is which magazines Dale Gribble and Comic Book Guy have in common. The hub ‘zines - those that unite groups with little else in common - provide a glimpse into the subjects that draw us together as Americans.
Inspired by the polarized political book networks on Orgnet, the power of Perl and the help of Adi Agafitei, I spent some time spidering around Amazon and dropping the results into Pajek, a free data visualization program from Slovenia’s University of Ljubljana. The resulting networks show a handful of magazines that link us all - families and singles, women and men, liberals and conservatives, teens and boomers, Hummer-drivers and tree-huggers…you get the idea.
The network above shows all of the shortest paths from ‘Mother Earth News’ to ‘Special Weapons Assault Team: The Magazine for Prepared Americans.’ The hub is ‘Popular Mechanics’, which I spent hours studying as a teenager. Men who read Popular Mechanics love their home workshops, and what is ‘Family Handyman’ but ‘Mother Earth News’ without the green baggage? Men who read ‘Popular Mechanics’ also like to hunt, fish and tinker with cars. From hunting it’s not a big step to ‘Guns and Ammo’, and from there it’s only a tiny step to assault weapon porn.
Popular Mechanics is a huge hub among men’s magazines. But it’s not the biggest. That distinction goes to Wired. More later.
Posted: December 5th, 2005 under .
Comments: 3
Comments
Comment from skrivare
Time: November 13, 2007, 2:57 pm
Interesting. Especially your comment, Chris.
Comment from Valdis
Time: December 26, 2007, 8:01 pm
You found 9 steps, which does not sound like much… but in social distance it is very far. Complete strangers[maybe even different species!] are at opposite ends of a nine step path.
What’s interesting is who is separated by 1,2,3 step paths… the rest is just social noise.
Good effort!
Comment from Christopher Fahey
Time: October 29, 2006, 4:07 pm
You probably already know this, but Amazon’s “collaborative filtering” engine does not, in fact, make connections between products, it makes connections between users. The product connections are a by-product of finding users who match your profile of products you like. Whether or not one product is similar to another one is entirely determined by whether or not they have a lot of users/customers in common, not by any quality inherent to the product itself. This insight, while not explicit, is embedded in the diagram above, but I thought it was worth mentioning. Cheers!