Hey, and welcome to today's session on Social Network Analysis Part 2. In part 1 of Social Network Analysis, I spoke about network structure, that basically we took static social networks, but networks are never static, where we took snapshots of of social networks, and then analyze the structures, develop different measures on the network in order to understand its structure a little bit better. Today, we will talk explicitly about network dynamics. That means how networks themselves change and how they change inside. So if we are with our computational scientific methods, we are still, and Social Network Analysis is in the analysis part, in analytical part. So social network analysis is an analysis which is falling in between the empirical and theoretical on these both ends. But today, explicitly, we will go a little bit beyond the analysis itself, network structures sort of analysis. We'll still do some analysis on networks, but we will go actually also in the theoretical part there. So that already shows you that it's not as clear cut as it might be insinuated here by this graph. Some of these methods you can use to analyze, but you can also use the new-build theory the same as you can use some tools that you use for building theory in order to analyze data, for example. So today, we will do the theory on networks and see how networks also should evolve. So that brings us a little bit to this last part of this course which is the theoretical aspect of social science. Okay. So again, we will answer three questions. The first question is how do networks evolve and the first question was actually answered by the second and the third question. The third question is how are we can predict what kind of network will forms. So how are networks evolved. Well, different networks will form over time so that the network itself changes. Then the second answer to the question on networks evolve is how can we predict what will happen inside the network. So the network changes inside itself and some dynamic is happening on the network we would say. So the first thing is the network itself changes, the node in the linked changes, and then in the third part we will talk about dynamics that happen on the network. In order to get a better glimpse of that, I want to start by showing you this video of Nicholas Christakis and James Fowler from USE about a network evolution over 32 years. Have a look, so what this video essentially shows is that weight gain happens in clusters. Now there are two interpretations to that, on the one hand, and that was the big headline as the title of the study, is the spread of obesity. The headline was off like obesity is contagious, right? It's not contagious like you touch somebody, and then they get infected and they gain weight. No, it's contagious in a social sense. In a sense of social social contagion. So basically you join a group and this group has a tendency to be overweight, you will also start to gain weight as being part of this group. So this is one interpretation, the spread, the social contagious effect of obesity. The other interpretation could also be homophily. Remember homophily, what that is? That basically birds of a feather flock together, that means that it could also be that if you're obese, you tend to join a group of obese people. If you are a sporty person, or a very athletic person, you tend to join groups with very athletic people. These effects are very subtle to piece apart homophily and or contagion. Unfortunately, we don't have time to go into the detail, but you can see that what we clearly saw here is that weight gain happens in clusters and we saw two effects, what happens on the network. On the one hand, and that is our second point, is we saw a different network forming. We saw these clusters forming of obese and non-obese people, and second of all there was happening also something inside the network, especially if you think about it like contagion, then something spread, something diffuses on the network. These are the two ways network can evolve and do evolves. So let's dive into the second question a little bit deeper. What kind of network will form and we actually predict that? One of the benefits of science one is explaining things and the other one is predicting things. So can we predict what kind of network will actually form, and luckily for us there are some particular shapes of networks that tend to form very frequently. The role of them is similar to if you have ever taken a statistics class, there is some kind of distributions you often run into, the normal distribution, the bell, the Poisson distribution, the exponential distribution, right? It doesn't mean these are the only distributions out there. There's an infinite number of distributions or how things can be distributed, but very often we run into a normal distribution. That's why it's called normal because it shows up quite often, or an exponential distribution, or a power law distribution. There's a handful of distributions that are very useful because they often turn up. Now, in networks something very similar happens. There's a handful of network formations that occur actually very frequently. It doesn't mean that these are the only network. There's an infinite number of network constellations, but particularly in social science, I picked here,four network constellations that occur quite frequently, and it's useful to understand them. One is random networks, second one is called scale-free networks. Third, small world networks, that's a technical term, seriously. The last one is hub-and-spoke or a star network.
