Plain-English definitions of the terms that come up in data projects. Each one starts with the definition, then says why it matters. Links go to the page that covers the topic in depth.
Jump to: Change data capture · DAX · Data lake · Data mining · Data pipeline · Data warehouse · ELT · Encryption at rest · Encryption in transit · ETL · iPaaS · Key management · Lakehouse · Master data management · Predictive analytics · RBAC · Reverse ETL · Semantic layer · Single source of truth · Star schema
Change data capture (CDC)
Change data capture is a method of copying only the rows that changed in a source database (inserts, updates and deletes) instead of reloading whole tables. It usually reads the database’s transaction log. CDC keeps a warehouse or downstream app close to real time while putting little load on the source system. See data integration services.
DAX
DAX (Data Analysis Expressions) is the formula language used in Power BI, Analysis Services and Excel Power Pivot to define measures and calculations, such as year-to-date revenue or margin percentage. Well-written DAX measures, defined once in a shared model, keep numbers consistent across reports. Badly placed logic is a common cause of slow Power BI reports. See Power BI consulting.
Data lake
A data lake is low-cost storage that holds large volumes of raw data in its original format: tables, JSON, logs, images and files. It’s flexible and suits data science and archiving, but without governance it becomes hard to find or trust anything in it. Most reporting still runs on a data warehouse or lakehouse layer built on top. See data warehouse consulting.
Data mining
Data mining is the process of finding patterns, correlations and anomalies in large datasets, such as customer segments, products bought together or unusual transactions. The results are used to explain behavior or to build predictive models. Data mining depends heavily on clean, combined data. See data mining services.
Data pipeline
A data pipeline is an automated process that moves data from a source system to a destination, such as a warehouse, often cleaning or reshaping it along the way. Good pipelines run on a schedule or continuously, log every run, retry temporary failures and alert someone when something breaks. See data integration services.
Data warehouse
A data warehouse is a central database that stores cleaned, structured data from many business systems, organized for reporting and analysis rather than day-to-day transactions. It keeps history, combines sources and holds business rules in one place, so every report reads the same numbers. Common cloud platforms include Snowflake, BigQuery, Redshift and Microsoft Fabric. See data warehouse consulting.
ELT
ELT (extract, load, transform) is a data integration pattern that loads raw data into the warehouse first, then transforms it there using the warehouse’s own compute, often with dbt. It keeps a raw copy for reprocessing and suits modern cloud warehouses. Compare with ETL. See data integration services.
Encryption at rest
Encryption at rest is the protection of stored data (databases, files, backups) by encoding it so it can’t be read without the correct key. AES-256 is the common standard. It protects data if storage media, files or backups are stolen or exposed. It’s usually built into cloud platforms but must be switched on and checked everywhere data lands. See data security services.
Encryption in transit
Encryption in transit is the protection of data while it moves between systems, such as from an app to a database or a pipeline to a warehouse, typically using TLS 1.2 or higher. It stops anyone intercepting the connection from reading the data. Every connection in a data platform should use it, including internal ones. See data security services.
ETL
ETL (extract, transform, load) is a data integration pattern that extracts data from sources, transforms it in a separate tool or server, then loads the finished result into the destination. It’s useful when sensitive fields must be removed or masked before data lands anywhere. Tools include Azure Data Factory, AWS Glue and SSIS. Compare with ELT.
iPaaS
iPaaS (integration platform as a service) is a hosted platform for connecting applications, with prebuilt connectors, a visual workflow builder, scheduling and monitoring. It suits companies with many apps and moderate integration needs. High-volume or complex integrations may be better as custom code. See application integration services.
Key management
Key management is the secure creation, storage, rotation, use and retirement of encryption keys. In the cloud it’s handled by services such as Azure Key Vault, AWS KMS and Google Cloud KMS. Good key management keeps keys separate from the data they protect, limits who can use them, rotates them on a schedule and logs every use. See data security services.
Lakehouse
A lakehouse is a data architecture that adds warehouse-style tables, transactions and governance on top of data lake storage, usually using open table formats such as Delta Lake or Apache Iceberg. It aims to serve both BI reporting and data science from one platform. Microsoft Fabric and Databricks are common examples. See data warehouse consulting.
Master data management (MDM)
Master data management is the practice of keeping one accurate, agreed version of core business entities (customers, products, suppliers, locations) across all systems. It includes matching and merging duplicates, setting ownership and rules for changes, and sharing the “golden record” back to the systems that need it. See data integration services.
Predictive analytics
Predictive analytics uses historical data, statistics and machine learning to estimate what’s likely to happen next, such as demand next month, which customers may churn or which invoices may be paid late. Predictions are only as good as the data behind them and should always be tested against a simple baseline. See data mining and predictive analytics.
RBAC (role-based access control)
Role-based access control is a security model that grants access by role rather than by individual, so a “regional sales manager” role sees regional sales data and nothing else. People get access by being assigned a role. RBAC makes permissions easier to review, and is often combined with row-level security in warehouses and BI tools. See data security services.
Reverse ETL
Reverse ETL is the process of copying cleaned, combined data from a data warehouse back into operational tools such as a CRM, support desk or marketing platform. It puts insights like customer lifetime value or churn risk in front of the people who act on them, inside the tools they already use. See data integration services.
Semantic layer
A semantic layer is the part of a BI system that translates database tables into business terms, such as “revenue” or “active customer”, each with one agreed calculation. Every dashboard reads from it, so the same metric always returns the same number. In Power BI it’s the semantic model; in Looker it’s LookML. See business intelligence consulting.
Single source of truth
A single source of truth is one agreed, governed place where a company’s key data lives, so every report, dashboard and app reads the same numbers. It’s usually a data warehouse fed by pipelines from each system, with metric definitions held in a semantic layer. It ends the debate over whose spreadsheet is right. See single source of truth use case.
Star schema
A star schema is a data model with a central fact table (events such as orders or invoices, with numbers like amount and quantity) linked to surrounding dimension tables (customers, products, dates, locations). It’s the standard design for reporting because it’s fast to query and easy for BI tools and people to understand. See data warehouse consulting.
Want these terms working for your business, not just defined?
Book a free 30-minute data audit.