chapter two

2 Graph Data Engineering

 

This chapter covers:

  • The main challenges related to Big Data as input to machine learning
  • How to handle Big Data analysis with graph models and graph databases
  • The shape and the features of a Graph Database

The previous chapter highlights the key role played by data in a machine learning project. Training the learning algorithm on a larger quantity of high quality data increases the accuracy of the model more than fine tuning or replacing the algorithm itself. Greg Linden, who invented for Amazon the now widely used item-to-item collaborative filtering algorithm, in an interview about big data replied:

“Big data is why Amazon’s recommendations work so well. Big data is what tunes search and helps us find what we need. Big data is what makes web and mobile intelligent.” [Peadar Coyle, 2016]

2.1   Working with Big Data

2.1.1   Volume

2.1.2   Velocity

2.1.3   Variety

2.1.4   Veracity

2.2   Graphs in Big Data Platform

2.2.1   Graphs are valuable for big data

2.2.2   Graph for master data management

2.3   Graph Databases

2.3.1   Graph Data Management

2.3.2   Sharding

2.3.3   Replication

2.3.4   Native graph vs not native graph

2.3.5   Neo4j

2.4   Summary

2.5   References