So, we discovered that this dataset does conform to a loose definition of the small world or a six-degree hypothesis. The average distance between any two nodes is about six. Also, as we're working with a sample, it's possible that working with a complete graph may fill in some links and bring the nodes closer together.
We also had an interesting time looking at some visualizations. One of the important lessons that we learned was that more complicated isn't always better. Simple, perhaps even a little boring, graphs can sometimes answer the questions we have in a better manner.
However, we've barely scratched the surface of what we can do with social graphs. We've primarily been looking at the network as a very basic, featureless graph, looking at the existence of people and their relationships without digging into the details. However, there are several directions we could go in to make our analysis more social. For one, we could look at the different types of relationships. Facebook and other social platforms allow you to specify spouses, for example, it might be interesting to look at an overlap between spouses' networks. Facebook also tracks interests and affiliations using their well-known Like feature. We could also look at how well people with similar interests find each other and form cliques.
In the end, we've managed to learn a lot about networks and how they work. Many real-world social networks share very similar characteristics, and there's a lot to be learned from sociology as well. These structures have always defined us but never more so than now. Being able to effectively analyze social networks, and the insights we can get from them, can be a useful and effective part of our toolkit.
In the next chapter, we'll look at using geographical analysis and applying that to weather data.
