> For the complete documentation index, see [llms.txt](https://docs.rootcause.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.rootcause.ai/core-technologies/ontology.md).

# Ontology

The ontology is the foundation of RootCause. It provides a structured map of your enterprise data, aligning columns to **entity**, **time**, and **location** concepts. By standardizing how data from different sources connects, it removes ambiguity and makes downstream causal analysis reliable and explainable.

***

### Definition & Purpose

An ontology is more than just a schema. It is a unified model of your data that organizes columns from many datasets into consistent, shared concepts.

Different datasets rarely use the same schema or naming conventions. RootCause scans every dataset and detects when columns in different datasets represent the same concept — for example, `customer_id` in a sales table and `cust_id` in a support table both mapping to a single shared Customer ID concept. Datasets that share an identifier concept can then be joined on the matching values of that concept. Joins operate on aligned values (exact matching by default, with tolerance options for numeric and time keys), so two datasets link when they record the same identifier scheme.

The purpose of the ontology is to create a shared foundation across disparate systems so RootCause can reason about cause and effect in a consistent way.

***

### How Data Links Together (Entity / Time / Location)

RootCause aligns columns to common anchors:

* Entity: Identifier concepts detected across datasets (e.g. Customer ID, Product SKU, Order Number) serve as join keys that link records referring to the same real-world object.
* Time: Temporal columns (timestamps, dates, periods) are classified as Time concepts so events can be ordered and aligned for time-series analysis.
* Location: Geographic columns (e.g. city, region, postal code) are classified as Location concepts, enabling location-based grouping and filtering.

By mapping columns into shared concepts, RootCause brings heterogeneous datasets into one coherent representation. Related concepts can also be organized into ontology groups — comparable measurements that share a unit, categorical variants of the same field, or replicas of the same source column — so they are treated consistently in analysis.

***

### Editing & Oversight

Ontologies aren't static. RootCause gives you tools to:

* Review and adjust concept mappings and classifications.
* Validate detected joins to catch errors or ambiguities early.
* Apply relationship rules — value constraints attached to concepts, sourced from domain knowledge, statistical detection, schema structure, user input, or calendar logic — so business-critical relationships stay correct.

This layer of oversight ensures data quality before it flows into causal graphs or simulations. For a hands-on walkthrough, see [Build Ontology](/user-guide/ontology-concepts.md).

***

### Data Views

RootCause generates Data Views — combined tables that reflect the joins across entities, times, and locations. A view that consolidates everything known about one identifier concept is called a 360 Table (e.g. a Customer 360 Table).

* Each view consolidates records into a single table that can be exported as CSV, JSON, or Parquet.
* Views are resolved dynamically from the ontology at query time — they are definitions, not materialized pipelines.

These Data Views form the structured inputs for causal analysis and Digital Twin simulations.