Summary

This chapter has taken you on a rollercoaster ride through how to develop knowledge graphs with the powerful igraph package. Firstly, we delved into the data preparation phases of knowledge graph construction, by looking at separating each abstract and then saving this into a separate cleaned abstract file.

Moving on, we looked at the steps needed to design the graph schema in the right way. This involved using popular NLP libraries such as spacy, plus a package we downloaded, and pip installed, the scispacy library for biomedical NLP tasks. Following this, we looked at extracting terms from our dataset and setting bounds on the frequencies of entities to include or exclude.

Once we had the foundations in place, we swiftly moved on to constructing a knowledge graph, from the ground up. This involved performing many of the key data modeling tasks we have been looking at in the chapters up until now. Furthermore, we made sure the graph contained the abstracts and terms...