Pipeline Development

This chapter will involve you, as a progressing graph data scientist, getting directly involved in building production-grade schemas. Here, we will teach you everything we have acquired from our years of experience as graph practitioners.

The use case for our pipeline design in this chapter will be to develop a schema that can be used to look at customers purchasing habits, with the ultimate aim of building a recommendations system that can be used as new (unseen) data is added to the graph. This will function very much like a streaming service, where, instead of You might like this film recommendations, you will be given recommendations on products you are likely to buy. We will look at querying methods looking at product similarity, alongside a popular similarity matching method called Jaccard similarity.

Again, you will be working extensively with Neo4j and Python to integrate and build the pipeline seen in many production environments. I hope you are...