How to be a good gossip…or at least a good computational modeler
Right now I’m writing to you and a couple seconds ago I was talking to my boss. Later this afternoon I have to write an email to someone on the other side of campus and perhaps my mom will call me sometime this evening. If you were to take a snapshot of the activity in our social network at one point in time:
…it will look a lot different in just a few seconds:
This isn’t much to get excited about, but it turns out it’s really hard to model this phenomenon computationally. Up until just a few years ago, network scientists have had to rely on the aggregates of these snapshots to estimate the system. Now, with more time-stamped data like mobile phone records and Twitter tweets, researchers have the data they need to start modeling temporality.
In a paper recently published in the physics arXiv, researchers from Northeastern professor Alessandro Vespignani’s lab used phone data from 6 million individuals (about 600 million calls!) to build a more robust model that not only takes into account temporality, but also deals with system’s history, which has a huge impact on the way information travels through it.
I tend to call my fiancé a lot more frequently than I call the kid I took to the semi formal in eighth grade (although I still think fondly of you, Johnny Depp). So the connection between Rob and me is a lot stronger than the one between Johnny and me. If you don’t consider the “memory” of our social interactions, the network looks a whole lot different…
…than if you do:
In the first case there are a lot of lines between the different nodes (people), representing dozens of paths along which information can travel. In the second there are only a few lines and a lot of unconnected nodes. If you drop a piece of information into the middle of the network (i.e., me trying to spread the rumor that I actually got to go on a date with Johnny Depp in middle school), it’ll have an easier time diffusing through the first system than through the second. But the real world looks nothing like the first system. Instead it looks like billions of the second one piled on top of and next to each other.
The colors in these last two images represent different kinds of people. Blue dots are for people that are “ignorant” of some piece of information. The red dots are for people who have heard the story from one person and pass it on to their neighbors (gossips!). The yellow dots get told the story and then deem it not interesting enough to pass it along anymore. As you can see there are a lot more gossipers in the unrealistic network than in the realistic one. In fact, in the more realistic network, it looks like information never goes anywhere, it keeps getting squelched by Debbie Downers who don’t think it’s amazingly awesome that I got to slow dance with Johnny Depp.
Well, it turns out it’s actually pretty difficult to spread rumors through tightly knit systems, according to the paper, of which post doctoral research associate Márton Karsai was the lead author. “It’s counterintuitive because you would think that your stronger ties would help spread information because you’re communicating a lot,” he explained. But in fact it isn’t until you tell someone whom you’re only weakly connected with that the rumor goes viral, so to speak. That’s because everyone in your tight little network already knows the information (or knows it’s not true), so they don’t feel compelled to pass it along. If you tell someone you’re only weakly connected with, it’s likely they’ve never heard this story before so they’ll be more likely to tell someone in their own tight little network. Then it spreads pretty quickly, spilling from one tight little network into the next.
This work was reported in a few mainstream media outlets over the last week or so, and not surprisingly: who wouldn’t want to know how to spread information if not journalists? But Karsai said the story is much bigger than the headlines were giving it credit for (“How Strong Social Ties Hinder the Spread of Rumours”). What’s really cool about this paper is that it’s the first one to ever successfully model how memory effects the spreading of information in human social network.
They’re data showed that a system’s memory, or the degree to which the people are linked, determines how far information will travel. If you take into account the temporality of a system (the fact that we’re not all talking to each other all the time), determines how quickly the information will spread.
But what’s a model without validation? Not much. Remember those 600 million phone calls they examined? It turns out their models did a much better job mimicking that real world data when they included memory and temporality in the mix.