Skip to main content

Data models

This section introduces the Cognite Data Fusion (CDF) data modeling features you can use to build expressive and scalable data models, ingest data to populate the models and query the models for the data they contain.

A data model is an abstract representation of real-world entities. It organizes data elements and their properties and standardizes how they relate. You can build a comprehensive, contextualized knowledge graph of your industrial data with a few simple building blocks—nodes, edges, and properties.

Cognite data models​

Data models and data storage

Cognite provides out-of-the-box data models to get you started quickly:

  • The core data model offers standardized building blocks, forming the basis for more specialized models.

  • The industry data models extend the core model to cater to specific industry requirements, for example for the processing industry.

  • Custom data models offer a data perspective tailored explicitly to a use case, solution, or application.

Data storage and analysis​

Containers are the physical storage for properties that logically belong together, such as sensor readings of temperature and pressure and equipment information like model, age, and maintenance history. For efficiency, certain types of industrial data—for example, files and time series data points— are kept in specialized data stores.

Views are a virtual layer for querying and selecting data in the containers for analysis, such as machine performance and predictive maintenance. Using the same data points allows you to create custom views to meet various needs while maintaining consistency and reliability.

Imagine that you need to analyze data from a factory assembly line. Your data model is the blueprint for the analysis, with views and containers defining how to store and access data efficiently and reliably. The data model could have these views and containers:

  • The Machine performance view retrieves last month's data from the Sensor readings and Production data containers to analyze how the equipment performs.

  • The Predictive maintenance view draws data from the Sensor readings container, focusing on readings that might indicate wear and tear. Also, it retrieves data from the Equipment information container, like maintenance history and machine age, to identify potential equipment failures.

More information​

Last updated