liveBook · Manning

Data model

As explained in Graph Neural Networks in Action

A data model is a formal structure that specifies the entities, attributes, relationships, and constraints of a dataset or information system.

A data model organizes a dataset for analysis by specifying its domain structure, schema, and usable instances. In graph workflows, it determines how entities and relationships become nodes, edges, and properties. section 8.2 section 8.3.1

Build and validate a model through a short practical pipeline:

  • Understand the domain, use case, relationships, and analytical questions. section 8.2.1
  • Create a data model, schema, and instance model that represent the relevant structure. section 8.2
  • Test the model against its schema and instances, then refactor it when reliability, performance, or accuracy requires changes. section 8.2

A graph data model also guides data exchange. Before choosing JSON, CSV, XML, or a language-specific format, establish the model, schema, storage structure, and receiving system’s expectations. section 8.3.1

Recruitment referrals

A recruitment system can model candidates and their referrals as an interconnected graph. The model should reflect domain questions such as optimizing referrals and identifying structures that govern candidate connections. section 8.2.1

Figure 8.2 Process of creating a robust graph data model from nongraph data
figure

Use the process diagram to trace the transformation from nongraph data to a tested and refined graph data model. section 8.2

How Graph Neural Networks in Action covers it

How deep it goes
Explains a practical design, testing, and refactoring workflow, while deferring detailed implementation of individual graph storage structures.
Examples
Uses recruitment referrals, JSON representations, candidate datasets, graph datasets, and published semantic models as concrete artifacts.
What it assumes
Assumes access to domain data and use-case questions, while establishing graph modeling from schema and instance-model basics.
What you can do after
Build it from scratch by defining, testing, and refactoring a graph model for a domain-specific analytical workflow.
How it connects
Connects domain knowledge, schemas, instance models, data structures, serialization, ETL, preprocessing, and semantic models.
Cover of Graph Neural Networks in Action
liveBookmanning.com
Read Graph Neural Networks in Action on liveBook Get the book at manning.com

Also explained in

Compare how 23 books explain data model →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Graph Neural Networks in Action. The section links point to the parts of the book it draws on.