CDDP key terms

The system of decision rights and accountabilities over data assets, defining who may take which actions, under what rules, and how compliance is monitored across the organization.

The operational practice of assigning named individuals to care for specific data domains, resolving quality issues, approving definitions, and enforcing governance policies day to day.

An accountable role with authority to approve access, classify sensitivity, and accept risk for a defined data domain or asset, usually at a senior management level.

The principles guiding fair, transparent, and responsible use of data, including consent, harm avoidance, bias mitigation, and accountability for algorithmic decisions affecting people.

Systematic error or unfair skew in collected data or model outputs that disadvantages groups, often arising from sampling, historical inequity, labeling choices, or proxy variables.

The process of assigning data to sensitivity tiers such as public, internal, confidential, or restricted, so that handling, access, and retention controls match the risk involved.

Documented rules specifying how long each data category is kept, where it resides, and how it is destroyed, balancing legal obligations, business needs, and privacy minimization.

The coordinated set of practices covering creation, storage, use, sharing, archiving, and deletion of data, ensuring each stage has defined controls and ownership.

A technology-independent diagram of major business entities and their relationships, used to agree on scope and meaning before physical design begins.

A refined model specifying entities, attributes, keys, and relationships without committing to a particular database product, bridging business concepts and physical implementation.

The implementation-level design of tables, columns, indexes, partitions, and constraints tuned for a specific database engine and workload.

Structuring relational tables to reduce redundancy and update anomalies by organizing attributes into progressively stricter normal forms based on functional dependencies.

Deliberately introducing redundancy or precomputed values into a schema to speed reads and simplify queries, accepting extra storage and update complexity.

A dimensional model with a central fact table of measures linked to surrounding dimension tables, optimized for analytical queries and business-friendly navigation.

A dimensional model in which dimension tables are normalized into related sub-dimensions, reducing redundancy at the cost of more joins in analytical queries.

A visual notation showing entities, their attributes, and the cardinality of relationships between them, commonly used to communicate data model designs.

Extract, transform, load: a pipeline pattern that pulls data from sources, applies cleansing and business rules, then writes results into a target store.

Extract, load, transform: a pattern that loads raw data into a target platform first and performs transformations there, leveraging scalable warehouse or lakehouse compute.

An automated sequence of steps that moves and transforms data between systems, including scheduling, error handling, monitoring, and dependency management.

Combining data from multiple sources into a consistent view, using patterns such as consolidation, federation, or propagation to serve analytical and operational needs.

A technique that presents a unified query layer over disparate sources without physically copying the data, trading some performance for freshness and reduced duplication.

An architecture combining cheap object storage and open table formats with warehouse-style transactions, schema enforcement, and governance to support both analytics and machine learning.

A centralized repository optimized for analytical querying, holding integrated, time-variant, subject-oriented data loaded from operational systems through controlled pipelines.

A subset of a warehouse scoped to one business area or team, providing focused, pre-shaped data for reporting with simpler access controls.

A large store holding raw and refined data in native formats on cheap storage, typically schema-on-read, serving diverse analytical and machine-learning workloads.

Measurable characteristics used to assess fitness for use, commonly accuracy, completeness, consistency, timeliness, uniqueness, and validity.

Statistical examination of datasets to reveal structure, value distributions, null rates, key candidates, and anomalies, forming a baseline for quality rules and modeling decisions.

Correcting or standardizing invalid, duplicate, incomplete, or malformed records so that data meets agreed quality rules before use in analysis or operations.

The discipline and tooling for creating authoritative, shared records of core entities such as customers, products, and suppliers, and synchronizing them across systems.

Controlled code sets and lookup values, such as country codes or currency lists, that classify and constrain other data and must be consistent enterprise-wide.

A recorded trace of where data came from, how it was transformed, and where it flows, supporting impact analysis, debugging, audit, and regulatory evidence.

Summary measures that characterize a dataset, including central tendency, dispersion, frequency, and shape, used to understand distributions before deeper analysis.

Methods that use sample data to draw conclusions about a wider population, including confidence intervals, hypothesis tests, and estimates of uncertainty.

A result unlikely to have occurred by chance under a stated null hypothesis, conventionally judged against a chosen probability threshold such as 0.05.

A measure of the strength and direction of association between two variables, ranging from minus one to one, which does not by itself establish causation.

Distortion arising when the selected sample is not representative of the target population, causing estimates and conclusions to systematically misstate reality.

Encoding quantitative and categorical information into visual forms such as charts and maps so patterns, comparisons, and outliers can be perceived quickly and accurately.

The practice of arranging metrics, filters, and supporting visuals on a single screen so a defined audience can monitor performance and act without extraneous clutter.

A key performance indicator: a small set of measures tied to strategic objectives, with agreed definitions, targets, and owners used to judge progress.

Combining visuals, narrative, and context to guide an audience from evidence to a recommended decision, emphasizing clarity of message over decorative presentation.

Protecting personal information through lawful collection, purpose limitation, consent, minimization, and honoring individual rights such as access and erasure.

Replacing sensitive values with realistic substitutes or redactions so non-production users and analysts can work without seeing identifiable or confidential content.

Substituting a sensitive value with a non-sensitive token mapped through a protected vault, reducing exposure while allowing authorized re-identification when required.

Granting permissions according to job roles rather than individuals, simplifying administration and enforcing least privilege across systems and datasets.

Protecting stored data by encoding it with cryptographic keys so that stolen disks, backups, or storage volumes remain unreadable without authorized key access.

Managed, elastically scalable services for storage, compute, and analytics delivered over the internet, billed by consumption and configured through provider controls.

An architecture using metadata, virtualization, and automated integration to connect distributed data sources into a coherent, governed access layer.

A socio-technical approach decentralizing analytical data ownership to domain teams who publish governed, discoverable data products against shared platform standards.

Continuously produced event records processed incrementally as they arrive, enabling near-real-time monitoring and reactions rather than periodic batch loads.