Summary 

In this chapter, we refreshed concepts such as graphs, nodes, and edges. We reviewed graph representation methods and explored how to visualize graphs. We also defined properties that are used to characterize networks, or parts of them.

We went through a well-known Python library to deal with graphs, networkx, and learned how to use it to apply theoretical concepts in practice.

We then ran examples and toy problems that are generally used to study the properties of networks, as well as benchmark performance and effectiveness of network algorithms. We also provided you with some useful links of repositories where network datasets can be found and downloaded, together with some tips on how to parse and process them.

In the next chapter, we will go beyond defining notions of ML on graphs. We will learn how more advanced and latent properties can be automatically found by specific ML algorithms.