Running algorithms

Raphtory implements many of the standard algorithms you expect within a graph library, but also has several temporal algorithms such as temporal_reachability and temporal motifs.

Raphtory categorizes algorithms into graphwide and node centric:

graphwide: returns one value for the whole graph (e.g. directed_graph_density, global_clustering_coefficient)
node centric: returns one value for each node in the graph (e.g. pagerank, degree_centrality)

Looking for a specific algorithm? Check out the Algorithm Library for a complete list of all available algorithms, including centrality measures, community detection, path finding, and temporal analysis.

For these examples we are going to use the One graph to rule them all dataset, which maps the co-occurrence of characters in the Lord of The Rings books.

This dataset is a simple edge list, consisting of the source character, destination character and the sentence they occurred together in (which we use as a timestamp). The dataframe for this can be seen in the output below.