concept graphs in category graphx

This is an excerpt from Manning's book Spark GraphX in Action. Login to get full access to this book.

GraphX is the graph processing layer on top of Spark that brings the power of Big Data processing to graphs—graphs that would be too large to fit on a single machine. People started using Spark for graphs long ago, including with the predecessor Bagel module, but with GraphX we now have a standardized way to do so, and it also provides a library of useful algorithms.

Figure 1.7. Different types of data that can be represented by graphs

The graphs we have looked at so far have captured links between pages; there is either a link or no link. We can make the graphs richer if we have more information about the connection. A typical example would be ratings information. When you give a 5-star rating to a movie on Netflix, not only do you create a connection between yourself and the movie, you also assign a value to that connection.

Movie ratings aren’t the only value that can be applied to connections in graphs. Dollar values in the analysis of financial fraud, distances travelled between cities, and the traffic carried across a network of mobile phone stations are other examples of ways to enhance the richness of the connections represented in graphs.