Network distribution and hubs

WANDER JAGER: The networks in society have grown over time. Initially, cities emerged in smaller agricultural areas, and you can see a small agricultural enterprise over here, “De Stadsakker” in Groningen. Today, we also see that the world is very connected, and agricultural hubs have emerged and are operating on a worldwide scale like, for example, this huge, sugar factory here in Groningen that ships worldwide. Some of these people that we call hubs have a global influence. For example, the president of China, pop stars, or the Dalai Lama have a worldwide impact. And we usually like to follow these people that are hubs. For example Katy Perry, Barack Obama have on Twitter more than 100 million followers.

Our tendency to follow these famous hubs is also mentioned preferential attachment, a very important concept in network theory. The NetLogo standard model on preferential attachment shows what type of network grows when new nodes prefer to be linked to nodes that already have more connections. This works just like kids wanting to befriend the more popular kids. Having just one friend more in the beginning of the group formation process makes you more attractive to befriend. And the more friends you get, the more other kids want to befriend you too. Becoming the popular kid on the block is thus also an emergent phenomenon.

You can see the degree distribution in the graph at the bottom of the screen, typically showing the fat-tailed power distribution we mentioned before where a lot of nodes have a few connections and just a few nodes have a lot of connections. These preferential attachment dynamics can be seen in different domains. A well-known example is the network of airline connections. As you know, there are many small airports and airstrips, but just a few hubs like Atlanta, Beijing, Dubai, and Heathrow London. For small airports, connecting to a hub is important, as it makes one-stop-over connections possible with many other airports. This makes you also more attractive for others to connect with. On social media, we also observe such preferential attachment dynamics.

Getting more followers on YouTube, Twitter, or Instagram, for example, makes you also more attractive for others to follow or support. This graph shows the distribution of followers on Instagram, and also here we see the typical, fat-tailed power distribution. Because these preferential attachment dynamics are important in understanding our social networks, we have developed an exercise in our simulated tribe. As you can see here, in our exercise model, we can also introduce hubs, and you can experiment with the impact of the number of hubs and how well they are connected to see how fast information travels through our simulated tribe in the following exercise.