data dictionary

A data dictionary is a controlled reference document or repository that defines the meaning, structure, allowed values, and usage rules for data elements within a system, interface, or dataset. In industrial and regulated manufacturing environments, it commonly documents how data fields are named, what they represent, how they are formatted, and how they should be used across OT/IT, MES, ERP, LIMS, QMS, and reporting systems.

Typical contents of a data dictionary

While the exact structure varies, a data dictionary for manufacturing and industrial systems commonly includes, for each data element or field:

• Unique name or identifier (for example, Batch_ID, Equipment_State)
• Business definition and description of what the data represents
• Data type and format (for example, integer, decimal, string, date-time, Boolean)
• Units of measure where applicable (for example, kg, °C, kWh)
• Allowed values or code lists (for example, enumerated states or status codes)
• Source system or originating equipment (for example, PLC tag, MES transaction, ERP module)
• Relationships to other data elements (for example, keys, references, hierarchies)
• Calculation or derivation rules for computed fields (for example, OEE components, KPIs)
• Ownership and stewardship (who maintains the definition)
• Version and change history, especially in regulated environments

Role in industrial and regulated environments

In manufacturing operations, a data dictionary helps ensure that different teams and systems interpret data consistently. It is often used to:

• Align MES, ERP, SCADA, historian, QMS, and analytics systems on common field meanings
• Support integration and mapping between systems by clarifying how fields correspond
• Document definitions of KPIs and metrics, including ISO 22400 style terms where applicable
• Provide traceable documentation for audits and inspections related to data integrity
• Support governance processes such as change control, validation, and data stewardship

For example, when defining custom KPIs such as availability or quality rates, a data dictionary can document how each input is defined, how it aligns or differs from standardized terminology, and how the KPI is calculated in each system where it appears.

What a data dictionary is not

To prevent confusion, it helps to distinguish a data dictionary from related concepts:

• It is not an operational database or historian; it describes the data but does not store the live values.
• It is not a full data model or schema diagram, although it can reference those and use some of the same concepts.
• It is not a full ontology or semantic model; it focuses on field-level definitions and usage rules rather than broader knowledge structures.

Common confusion and related terms

Data catalog: A data catalog typically indexes datasets, reports, and data assets across an organization, often including business metadata and ownership. A data dictionary usually goes deeper at the field level, describing individual columns, tags, or attributes.

Business glossary: A business glossary defines business terms and concepts (for example, “batch”, “lot”, “work order”) in plain language. A data dictionary often maps those business terms to specific fields and structures in systems.

Use with standards and KPIs

When organizations align custom KPIs or data structures with standards such as ISO 22400, the data dictionary is a common place to document how each local field or metric relates to the standardized term. This can include noting:

• Which ISO 22400 concept a local field or KPI aligns with, if any
• Any intentional deviations from the standard definition
• How legacy MES/ERP/QMS fields map to standardized or harmonized names

Documenting these relationships in a data dictionary supports consistent use of terminology across systems, helps avoid ambiguity, and provides a reference point during system integration, validation, and audits.