Summary

In this chapter, we went through several different approaches to community detection. Each had its pros and cons.

We saw that connected components can be useful for identifying communities, but only if the network consists of more than just one single primary component. To use connected components to identify communities, there need to be some smaller connected components split off. It’s very important to use connected components at the beginning of your network analysis to get an understanding of the overall structure of your network, but it is less than ideal as a standalone tool for identifying communities.

Next, we used the Louvain method. This algorithm is extremely fast and can be useful in networks where there are hundreds of millions of nodes and billions of edges. If your network is very large, this is a useful first approach for community detection. The algorithm is fast, and the results are clean. There is also a parameter you can experiment with to...