Glossary · Information modeling
Data model
German: Datenmodell
In information architecture, a data model is a structured definition of data types, fields, relationships, constraints and meaning. It describes what data exists and how its parts relate, independent of any single message or file.
- Information architecture
- Ontology
- German source term
In one sentence
A data model defines data types, fields, relationships, constraints and meaning: what data says, as opposed to the protocol that carries it.
Example
The data model of a robot cell defines a Job with an ID, a recipe, a target quantity in pieces and a status from a fixed list, and relates each Job to the products it produces.
How it applies
- Integrated systems: Two systems can exchange data meaningfully only if they share a data model, not just a connection. Define units, allowed values and meaning for each field, and version the model (data model version).
- Technical documentation: Content has data models too: iiRDS metadata and DITA information types define what a topic is and how it relates to products and components.
- AI and retrieval: A documented data model tells AI pipelines what fields mean, which reduces misinterpretation when data is used to answer questions or generate text.
Data model vs. protocol
A protocol defines how data is transmitted: message format, sequence, timing and error handling. A data model defines what the data means. The same data model can be carried over different protocols, and one protocol can carry different data models.
Data model vs. information model
An information model is technology-neutral and describes concepts and rules. A data model is closer to implementation, and a schema is its formal, checkable expression.