Summary

In this chapter, you learned about the GDS Python client. From graph management (projection, retrieval, and deletion), to running algorithms and retrieving their results in a pandas dataframe, and all we have done in the preceding chapters with Cypher, you are now able to do it without needing to open the Neo4j browser anymore. By only using a Jupyter notebook, you can take advantage of the full power of Neo4j and the GDS. Since the GDS procedures return pandas dataframes, it is quite straightforward to include these results within a Python ML pipeline, for instance, by using scikit-learn, as we have done in the last section of this chapter.

This chapter and the preceding ones have shown you how to extract features from a graph dataset, taking advantage of the graph structure. Features such as a degree, or more generally, centrality metrics, and community ID are only available if you consider the relationships between the entities in your dataset to build a graph. Depending...