ICCP CDP key terms

The practice of governing data from creation through archival or destruction, defining policies for each stage so that data remains accurate, available, and compliant throughout its useful life.

The process of examining a dataset to collect statistics and patterns about its structure, completeness, value distributions, and anomalies, typically performed before designing ETL or assessing data quality.

The activity of detecting and correcting or removing inaccurate, incomplete, duplicated, or malformed records so that data meets agreed quality standards for its intended use.

Measurable characteristics used to assess fitness for use, commonly including accuracy, completeness, consistency, timeliness, uniqueness, and validity, each evaluated against defined business rules.

A discipline and set of processes that create and maintain a single, trusted, shared view of core business entities such as customers, products, and suppliers across systems.

A controlled set of permissible values or classification codes, such as country codes or product categories, used to standardize data across applications and enable consistent interpretation.

Documentation describing the structure of a dataset, including entity names, attribute definitions, data types, formats, allowed values, and relationships, serving as a shared reference for developers and analysts.

Structured information that describes other data, such as its origin, meaning, format, ownership, lineage, and usage, enabling discovery, interpretation, and management of data assets.

A record of the origin and movement of data across systems, transformations, and reports, showing where values came from and how they were changed along the way.

Extract, Transform, Load: a data integration pattern that pulls data from source systems, applies cleaning and transformation rules, and writes the results into a target warehouse or store.

Extract, Load, Transform: a pattern that loads raw source data into a target platform first and applies transformations there, typically leveraging the scalable compute of modern cloud warehouses.

A centralized repository that integrates data from multiple operational sources, structured for query and analysis and optimized to support reporting and decision-making rather than transaction processing.

A subset of a data warehouse, scoped to a particular business area or user group, that presents integrated data in a form tailored to that group's analytical needs.

A modeling technique that organizes analytical data into fact tables containing measures and dimension tables containing descriptive context, designed for understandable and performant querying.

A dimensional design in which a central fact table references surrounding dimension tables through foreign keys, producing a simple, denormalized structure well suited to business intelligence queries.

A dimensional design in which dimension tables are normalized into multiple related tables, reducing redundancy at the cost of more joins and greater query complexity.

The central table in a dimensional model that stores quantitative measures and foreign keys to dimension tables, typically at a defined grain such as one row per transaction or per day.

The level of detail represented by a single row in a fact table, declared explicitly during design so that measures are neither double-counted nor aggregated too coarsely.

A technique for managing dimension attributes that change over time, using methods such as overwriting values or retaining history with versioned rows and effective dates.

A multidimensional data structure that pre-aggregates measures across dimensions, allowing analysts to slice, dice, drill down, and pivot data quickly for analytical exploration.

The overall system of decision rights, policies, and accountabilities that determines how data is collected, managed, protected, and used across an organization.

The operational execution of data governance, in which designated individuals apply standards, resolve quality issues, and maintain definitions for a defined data domain.

The assignment of accountability for a data domain to a role or person who approves definitions, access decisions, and quality targets for that data.

A role responsible for the technical storage, movement, backup, and security of data on behalf of data owners, without necessarily deciding what the data means or who may use it.

A formal statement of rules and principles governing data handling, such as retention, access, privacy, and acceptable use, that guides decisions and behavior across the organization.

Rules specifying how long each class of data must be kept, where it is stored, and when and how it must be destroyed, balancing legal obligations with business needs.

The process of assigning data to categories such as public, internal, confidential, or restricted, based on sensitivity and regulatory requirements, to determine appropriate handling controls.

A technique that substitutes sensitive values with realistic but altered data, either permanently or dynamically at query time, so that users can work without seeing protected information.

The set of practices and obligations protecting personal information from unauthorized access or misuse, including consent management, minimization, and honoring individual rights over their data.

An organized inventory of data assets with searchable metadata, definitions, ownership, and lineage, helping users find, understand, and trust available data.

An investigative approach that uses summary statistics and visualizations to reveal distributions, relationships, and anomalies before formal modeling or hypothesis testing.

The use of statistical or machine learning methods to estimate an unknown outcome from available variables, supporting classification, ranking, or forecasting decisions.

A machine learning approach that trains a model on labeled examples, learning a mapping from input features to a known target so it can predict that target for new cases.

A modeling failure in which a model fits training data so closely that it captures noise and performs poorly on new data, often detected by comparing training and validation performance.

A resampling technique that partitions data into multiple folds, training and validating on different splits, to estimate how well a model will generalize to unseen data.

The representation of data through charts, graphs, and dashboards so that patterns, comparisons, and outliers are communicated clearly to an intended audience.

Key Performance Indicator: a quantifiable measure tied to a strategic objective, chosen so that progress toward that objective can be monitored consistently over time.

The application of statistical and computational techniques to large datasets to discover previously unknown patterns, associations, or segments that inform business decisions.