In this chapter, we have learned about the different ways to build graphs in Spark by working with concrete examples borrowed from online social networks, food science, and e-mail communications. We have seen that constructing a graph requires some data preparation and wrangling efforts. Nonetheless, GraphX offers various graph builder functions from which we can choose, depending on the graph representation that we need to create, and on the shape of the available datasets. Such usable functionalities are the advantages of GraphX against other similar graph-processing frameworks. Moreover, we looked at some basic statistics and properties of graphs, which are rather useful in characterizing their structure and in understanding their representation.

In the next chapter, we will go deeper into the analysis of graphs, using data visualization tools and new graph-theoretical concepts and algorithms, such as connectedness, triangle counting, and PageRank.